The Compensatory Effects of Pictorial and Verbal Information

THE COMPENSATORY EFFECTS OF PICTORIAL AND VERBAL INFORMATION

FOR HAPTIC INFORMATION ON CONSUMER RESPONSES

IN NON-STORE SHOPPING ENVIRONMENTS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Minjung Park, M.A.

*****

The Ohio State University

2006

Dissertation Committee:

Professor Sharron J. Lennon, Adviser Approved by

Professor Leslie Stoel ______

Professor Jay Kandampully Adviser

Professor Michael Browne College of Human Ecology

ABSTRACT

The purpose of the study was to examine the compensatory effect of pictorial and

verbal information for haptic information in non-store shopping environments. Dual coding theory and Stimulus-Organism-Responses paradigm provided the theoretical

frameworks for the study. The proposed model of the study was examined by conducting

an experiment using a mock apparel website and a mock apparel catalog. Additionally,

this research addressed an alternative model based on the results of the originally

proposed model testing. Multivariate analyses of variance and structural equation

modeling were used to test both the originally proposed model and the alternative model.

The design of the originally proposed model was a 2 (picture swatch or no

swatch) x 2 (high haptic imagery description or low haptic imagery description) x 2

(online shopping or catalog shopping) x 2 (high need for touch or low need for touch)

between-subjects factorial design. Dependent variables for the original model were

perceived product quality, perceived risk, attitude toward a product, and purchase

intentions. The alternative model used the same independent variables and dependent

variables with the originally hypothesized model and added two more dependent

variables (perceived haptic imagery, perceived interactivity).

The findings from the original model and alternative model revealed: (1) the main

effect of pictorial information on perceived haptic imagery; (2) the main effect of verbal information on perceived haptic imagery; (3) the main effect of shopping contexts on perceived interactivity; (4) the positive relationships between perceived haptic imagery and perceived product quality; (5) the positive relationships between perceived interactivity and perceived product quality and the negative relationships between perceived interactivity and perceived risk; (6) positive relationships between perceived product quality and attitude toward a product, between perceived product quality and behavioral intentions, and between attitude toward a product and behavioral intentions;

(7) negative relationships between perceived risk and attitude toward a product and between perceived risk and behavioral intentions.

The study provides theoretical and practical implications. The empirical evidence of the compensatory mechanism for haptic information contributes to the literature in non-store retailing and apparel retailing fields and provides effective marketing strategies using judicious information presentation.

iii

Dedicated to my husband

Jungho Nam

ACKNOWLEDGMENTS

I would like to express my sincere appreciation to all those people who provided me with assistance during my research. I am especially thankful to my adviser, Dr,

Sharron Lennon, whose constant encouragement, enthusiasm, and love enabled me to complete my graduate program and dissertation. Her intellectual support and extensive knowledge provided me with my learning opportunities to explore and to enhance a number of research perspectives. Words will never express my gratitude for the kindness, care and support she bestowed upon me.

I am grateful to Dr. Leslie Stoel who stimulated discussions and great ideas for my study. She facilitated additional opportunities for independent research and scholarly writings in the field of online apparel retailing. I wish to thank Dr. Jay Kandampully who provided valuable suggestions and various aspects of this dissertation and gave me opportunities to discuss my research perspectives. I am thankful to Dr. Michael Browne who taught me structural equation modeling and offered his continuous support of statistical issues regarding this dissertation. My appreciation is also extended to Dr. Sook-

Ja Lim for her advice and encouragement from my master’s program.

I wish to thank my friends for their friendship: Hyejeong Kim, Wi-Suk Kwon,

Sejin Ha, Minjeong Kim, Jung-Hwan Kim, Jiyoung Lim, Young Ha, Jungmin Yoo,

Hyunjoo Im, and Jisun Park. A special note of appreciation goes to my dearest friends,

Minjeong Kim and Jungmin Yoo. The inspiration and intellectual and emotional support

of Dr. Minjeong Kim fostered my hope throughout my graduate program and helped me get through difficult times. Jungmin Yoo was always there to share knowledge and opinions on research.

I reserve distinct gratitude for my husband, parents and parents-in-law who always give endless love and consideration to me. I am deeply grateful to my loving husband, Jungho Nam, who endlessly supported me to achieve my goal and provided me the energy I needed to finish my work. I could not have completed this study without his support and love. My parents and parents-in-law emotionally and financially supported me, and encouraged me never to give up and to always seek their wisdom and love. And finally, thank you to my sisters, Hyejung Park and Hoojung Park, brother-in-law, Woong

Jae Lee, and my sisters-in-law, Sang-Hee Nam, Sang-Min Nam, and Sang-Hyun Nam for their encouragement and support.

VITA

February 12, 1975 ………………………………….Born – Seoul, Korea

1999…………………………………………………B.A. Clothing and Textiles; B.A. Consumer Science & Human Development, Ewha Womans University

2001…………………………………………………M.A. Clothing and Textiles, Ewha Womans University

2002 - Present……………………………………….Ph.D. Textiles and Clothing, Department of Consumer Sciences, The Ohio State University

PUBLICATIONS

1. Park, M., & Lennon, S. J. (2006). Developing a conceptual model to explain the effect of site atmospherics on online shoppers’ responses [CD-ROM]. In D. J. Burns (Ed.), Spring 2006 ACRA Proceedings. Philadelphia, PA: Xavier University.

2. Park, M., Lennon, S. J., & Stoel, L. (2005). The roles of product and customer service information in determining website quality, satisfaction, and patronage intentions. Abstract published in Proceedings of the International Textiles and Apparel Association (No. 62).

3. Park. M, Stoel, L., & Lennon, S. J. (2005). Key dimensions of electronic service quality: applying qualitative research. Abstract published in Proceedings of the Seoul International Clothing and Textiles Conference.

4. Park, M., Stoel, L., & Lennon, S. J. (2005). Developing a conceptual model to explain the effect of information quality on website quality perceptions and customer satisfaction [CD-ROM]. Abstract published in D. J. Burns (Ed.), Spring 2005 ACRA Proceedings. Philadelphia, PA: Xavier University.

5. Park, M., & Lennon, S. J. (2005). The impact of service recovery and service failure on service quality, customer satisfaction and behavioral intentions in online vii

shopping. Abstract published in the First Virtual Conference of the Global Symposium for Consumer Sciences. Available: http://www.Consumersciences.org/abstract.html

6. Park, M., & Lennon, S. J. (2004). Self-consciousness, materialism, compulsive buying, and conspicuous consumption of clothing. Abstract published in Proceedings of the International Textiles and Apparel Association (No. 61).

FIELDS OF STUDY

Major Field: Human Ecology Area of Specialization: Textiles and Clothing Minor Field: Quantitative Psychology

viii

TABLE OF CONTENTS Page

Abstract...... i Dedication...... iv Acknowledgments...... v Vita...... vii List of Tables ...... xii List of Figures...... xv

Chapters:

1. Introduction...... 1 Problem statement...... 4 Purpose of the study...... 6 Significance of the study...... 7 Definition of terms...... 8

2. Literature review...... 10 Overview of non-store shopping...... 10 Characteristics of online and catalog shopping ...... 12 Information presentation...... 20 Theoretical framework...... 26 Dual coding theory and imagery...... 26 Stimulus-Organism-Response (SOR) paradigm...... 33 Hypothesis development...... 36 The effect of environmental cues on internal states...... 38 The moderating effect of individual characteristics...... 45 The relationships between internal states and consumer responses...... 46

3. Pretests ...... 49 Pretest 1...... 51 Pretest 2...... 54 Pretest 3...... 57 Pretest 4 (content analysis) ...... 59 Method ...... 59 Results...... 61 Pretest 5...... 67

4. Main experiment ...... 74 ix

Experimental design...... 74 Mock website and catalog development...... 77 Sample and procedure...... 78 Measures ...... 80 Perceived product quality ...... 81 Perceived risk...... 81 Attitude toward a product ...... 82 Behavioral intentions ...... 82 Need for touch (NFT) ...... 82

5. Analysis and results ...... 86 Sample characteristics...... 86 Preliminary analysis...... 92 Descriptive and factor analyses of measurements ...... 92 Initial model considerations...... 101 Manipulation checks ...... 109 Hypothesis testing...... 111 Part I (Analysis of Variance) ...... 112 Part II (Structural Equation Modeling)...... 115 Post hoc analysis (An alternative model)...... 123 Descriptive and factor analyses of measurements ...... 125 Initial model considerations...... 130 Alternative model testing...... 137

6. Discussion and conclusion...... 153 Summary and conclusion...... 153 The effect of pictorial and verbal information...... 155 Individual differences ...... 158 Situational differences ...... 159 The relationships among consumers’ internal states and shopping outcomes ...... 161 Implication ...... 165 Theoretical implications and contributions...... 165 Managerial implications and contributions...... 168 Limitation...... 171 Future research...... 172

Bibliography ...... 175 Appendices...... 188 Appendix A: Invitation mail and consent form ...... 188 x

Appendix B: Pretest1 ...... 191 Appendix C: Pretest 2 ...... 194 Appendix D: Pretest 3...... 197 Appendix E: Pretest 5 ...... 201 Appendix F: Main experimental conditions ...... 205 Appendix G: Questionnaire ...... 225 Appendix H: Human subject exemption form...... 232

LIST OF TABLES

Table Page

3.1 Ratings of Apparel Categories...... 52

3.2 Repeated measures MANOVA and ANOVA Results for Importance of Touch and

Fit by Apparel Categories ...... 53

3.3 Paired Sample t-Test Results for the Comparisons among Apparel Categories in

Importance of Touch and Fit...... 53

3.4 Ratings of Apparel Stimuli ...... 56

3.5 t-tests for the apparel stimuli and models ...... 57

3.6 Effects for visual information (picture swatch vs. no swatch) on vividness and

haptic imagery elaboration...... 59

3.7 Verbal and Pictorial Information Coded from Apparel Websites and Catalogs....65

3.8 Rating for descriptions of dresses...... 68

3.9 Effects for verbal information (haptic imagery vs. non haptic imagery

descriptions) on haptic imagery related to material properties...... 71

3.10 Effects for verbal information (haptic imagery description vs. non haptic imagery

description) on haptic imagery...... 73

4.1 Manipulated experimental conditions for the study...... 76

4.2 Items for dependent variables ...... 84

5.1 Participants for the Experimental Conditions...... 87

xii

5.2 Sample Demographic Characteristics...... 89

5.3 Participants’ Online and Catalog Shopping Experience...... 91

5.4 EFA Results for the 10item, 3-Factor Model for Perceived Product Quality...... 95

5.5 EFA Results for the 12item, 4-Factor Model for Perceived Risk...... 98

5.6 Descriptive Statistics and Reliability for Four Variables ...... 100

5.7 Results from EFA of the Finalized Measurements...... 105

5.8 Results from CFA of the Finalized Measurements...... 106

5.9 Correlation Coefficients and Confidence Interval for Discriminant Validity .....108

5.10 Chi-square Difference Test for Discriminant Validity ...... 108

5.11 t-test Results for the Manipulation Checks ...... 110

5.12 Summary of H1 through H5 ...... 112

5.13 MANOVA Results for Hypotheses Testing (H1 through H5)...... 114

5.14 ANOVA Results for the Comparisons between High NFT and Low NFT

Groups...... 115

5.15 Results from the SEM for Testing H6 through H10...... 122

5.16 Decomposition of Direct, Indirect, and Total Effects for the Hypothesized

Model ...... 123

5.17 EFA Results for the 8item, 3-Factor Model for Haptic Imagery...... 128

5.18 Descriptive Statistics and Reliability for the Additional Measures...... 129

5.19 Results from EFA of the Added Measurements for the Alternative Model ...... 133

xiii

5.20 Results from CFA of the Finalized Measurements for the Alternative Model....134

5.21 Correlation Coefficients and Confidence Interval for Discriminant Validity .....136

5.22 Chi-square Difference Test for Discriminant Validity ...... 137

5.23 MANOVA Results for the Part I of the Alternative Model...... 139

5.24 ANOVA Results for the Effect of Pictorial and Verbal Information on

Haptic imagery...... 141

5.25 ANOVA Results for the Effect of NFT on Perceived Haptic Imagery and

Perceived Interactivity...... 142

5.26 Results from the SEM for Testing the Alternative Model...... 150

5.27 Decomposition of Direct, Indirect, and Total Effects for the Alternative

Model ...... 152

6.1 Summary of Results of the Hypothesis Testing...... 164

xiv

LIST OF FIGURES

Figure Page

2.1 An S-O-R model of consumer response to online shopping (Eroglu, Machleit, &

Davis, 001, p. 179) ...... 35

2.2 Research model: The effects of information on customer internal states and

shopping outcomes in non-store apparel shopping contexts...... 37

3.1 The procedure of the study ...... 50

5.1 A CFA Model for Final Measurement Items...... 103

5.2 Proposed Model of the Study ...... 111

5.3 A Structural Equation Model for Testing H6 through H10 ...... 117

5.4 The Alternative Model of the Study ...... 125

5.5 A CFA Model of Final Measurement Items for the Alternative Model ...... 131

5.6 Pictorial Information by Shopping Context on Interactivity ...... 144

5.7 Shopping Context by NFT on Vividness...... 144

5.8 A Structural Equation Model for Testing the Second Part of Alternative

Model ...... 146

CHAPTER 1

INTRODUCTION

Non-store shopping channels play a strategic role for retailers in today’s competitive market place. Online retail sales have dramatically increased. According to the U.S. Department of Commerce, U.S. online retail sales for the first quarter of 2005 were $19.8 billon, and the growth rate of online retail sales over the fourth quarter of

2004 reached 6.4% (U.S. Census Bureau News, 2005). Moreover, online retail sales are expected to reach $229.9 billion in 2008 (Cantwell, 2004). While total retail sales for the first quarter 2005 increased 7.3% from the first quarter of 2004, e-commerce estimates increased by 23.8% in the same period (U.S. Census Bureau News, 2005). Although the growth rate of online shopping has outpaced that of catalog shopping, catalogs still remain an effective means of access to consumers. According to U.S. consumer shopping surveys, 60% of Americans shop from catalogs each year (Grant, 2002) and 49.6% of shoppers made at least one catalog purchase in 2000 (Sherry, 2002). To capitalize on these trends, traditional in-store retailers are providing websites and catalogs. A survey of the best performing online retailers reported that among the list of the top 50 online retailers, 43 retailers operated multi-shopping channels in 2003, compared to only 27 in

1999 (Gallo & McAlister, 2003).

Compared to in-store shopping, non-store shopping provides different

characteristics in terms of physical environments. Non-store shopping allows consumers

to shop from worldwide stores and to save time associated with information and store

searching tasks in traditional shopping (Koering, 2003; Ranganatha & Ganapathy, 2002).

Non-store shopping environments provide not only timesaving conveniences but also

economical benefits for consumers by offering relatively low price products (Eastlick &

Feinberg, 1999). Favorable and pleasant environments in which non-store shopping

occurs (i.e., at home, 24/7 convenience) attract consumers and allow them to efficiently

search for information and enjoy the shopping process (Stell & Paden, 1999). Extensive

product and service information provided by online and catalog retailers may help

consumers reduce uncertainty about evaluating a product and make better decisions (Kim

& Lennon, 2000; Park, Lennon, Stoel, 2005).

The importance of haptic information (i.e., sensory information obtained through

touch by hands) has been emphasized in non-store shopping environments (Peck &

Childers, 2003a). In particular, for apparel shoppers, touch can be an important

information source. Direct sensory contact with a fabric and a garment may provide apparel shoppers with valuable product information in order that they evaluate product

quality and decide on a purchase (Mooy & Robben, 2002). However, in the context of

non-store shopping, it is not possible to provide an opportunity for direct sensory

examination for consumers. Many consumers search for information via websites or

catalogs, however they tend to abandon actual purchases (Shim, Eastlick, Lotz, &

Warrington, 2001) because of the inability to physically examine a product. Since online

and catalog apparel shoppers can not immediately acquire products and directly try on or

touch a product, they are uncertain about product quality (Maignan & Lukas, 1997;

McCabe & Nowlis, 2003). This lack of direct sensory experience may make it difficult for apparel shoppers to assess product quality and subsequently increases their perceived risk in online and catalog apparel shopping.

Non-store retailers strive to overcome this intangibility dilemma (Koering, 2003).

Pictorial and verbal information are fundamental features of consumer information in the context of non-store shopping. Researchers maintain that because touch is unavailable in non-store shopping contexts, written descriptions and visual depictions of products may serve to compensate for haptic information (Peck & Childers, 2003a). For example, pictures (e.g., a picture of fabric swatch) and verbal descriptions (e.g., descriptions of fabric hand and fabric properties) associated with haptic information may help shoppers recall their memories of past purchases and experience in touching a similar fabric and perceive haptic information, enhancing consumers’ favorable evaluation of a product and reducing perceived risk.

The verbal and pictorial information might be differently processed according to differences in situations. As compared to catalog shopping environments, online shopping environments provide a unique feature which is interactivity. Interactive tools

(e.g., chat rooms and email) support exchanges of opinions between consumers and online retailers, and multi-media features (e.g., click function, hyperlinks, and zoom function) allow consumers to control and manage contents (e.g., colors and graphics).

Regardless of their physical locations, the interactive features strengthen the relationships with consumers (Ko, Cho, & Roberts, 2005). Compared to static and passive catalog shopping environments, consumers may process pictorial and verbal information

differently in the interactive online shopping environment. In addition, individual

differences in need for touch (NFT) should be considered in information seeking and

processing situations. For example, people who are highly motivated to directly examine a product (high NFT) are likely to seek haptic information and scrutinize quality and properties of a fabric as compared to people who have a low motivation to directly inspect a product (low NFT).

Although the importance of haptic information has been discussed, information processing research on diverse non-store shopping environments has been limited.

Therefore, the present study focuses on (1) investigating the potential compensatory effect of pictorial and verbal information for haptic information on consumer responses and (2) examining the effect of individual differences in need for touch (NFT) and situational differences (catalog apparel shopping environment vs. online apparel shopping environment) on consumer responses.

Problem Statement

The Internet and catalogs are effective tools of direct marketing and generate significant retail sales volume. Although benefits and obstacles of catalogs and the

Internet have been observed, there is a lack of empirical research comparing similarities and differences between these two non-store shopping channels. A frequently identified problem associated with shopping online and from a catalog is the inability to directly touch and physically examine a product (Mooy & Robben, 2002; Vijayasarathy & Jones,

2000). This problem might be more critical when online and catalog shoppers evaluate

apparel as compared to other product categories (e.g., computers, electronics, CDs, books). Previous research has found evidence that due to the lack of sensory experience, catalog and online apparel shoppers perceive risks associated with online and catalog purchases and may be reluctant to purchase a garment online or from catalogs (Park et al.,

2005). In spite of the significant importance of identifying and understanding possible compensatory mechanisms for touch information, a lack of attention to this issue has been recognized in the literature (Peck & Childers, 2003a; 2003b).

Research on risk reduction strategies in non-store apparel shopping has focused on the roles of quantity and quality of information, product presentations, and store atmospherics. Although elements of information, manner of product presentation, and store atmospheric cues have a potential to compensate for haptic information, surprisingly the underlying compensatory mechanisms have been addressed by very little research.

In particular, as fundamental elements of the information environment, pictorial and verbal information may be able to substitute for haptic information. The possible compensatory effect of pictorial and verbal information for haptic information in online and catalog shopping channels appears to be very interesting. Moreover, there is a lack of theoretical perspectives in terms of how pictorial and verbal information are activated as proxies of haptic information.

Another consideration of the compensatory process is the effect of situational and individual factors. As compared to catalog shopping environments, online shopping environments can facilitate information searching activities of shoppers by providing advanced interactive features (e.g., close-ups, zoom function). Although the superior

features of online shopping environments over catalog shopping environments have been

observed, little research has considered differences across the two shopping channels.

Depending on an individual’s preference for the use of haptic information, the

compensatory mechanism might vary. Recently, a few studies recognized the importance of individual differences in terms of a preference for the utilization of haptic information.

Purpose of the Study

This study focuses on addressing the research gap discussed in the previous

section. The purpose of the study is to investigate the compensatory effects of pictorial

and verbal information for haptic information in non-store apparel shopping contexts and

to examine individual and situational differences in information processing. This study

attempts to provide theoretical perspectives to explain the mechanism by which

information presentation and shopping contexts influence customer cognitive evaluations

and behavioral responses. Based on the Stimulus-Organism-Response (S-O-R) paradigm,

the model of this study was constructed, which explains that information presentation and

shopping contexts (S) have an impact on customers’ behavioral responses (R), through

customers’ cognitive evaluations (O). Dual coding theory was applied to understand that

pictorial and verbal information which evokes haptic imagery affects customers’

perceived product quality and perceived risk, subsequently influencing attitude toward a

product and purchase intentions.

Based on the reviews of literature and theoretical frameworks (Chapter 2), the specific purposes of the study were developed below:

(1) to investigate that as proxies for haptic information, pictorial and verbal information have an impact on customers’ internal states (perceived product quality, perceived risk) (2) to explore the effect of shopping contexts on perceived product quality and perceived risk (3) to explore if need for touch moderates the effect of pictorial and verbal information on consumer internal states (4) to investigate the relationships among perceived product quality, perceived risk, attitude toward a product, and purchase intentions

Significance of the Study

This study is expected to provide theoretical and practical implications in non-

store retailing and apparel retailing fields. Regarding the lack of direct and sensory

experience, the compensatory mechanism of pictorial and verbal information for haptic

information has not been thoroughly examined in literature. The empirical evidence of

this study will contribute to the literature in information processing and non-store apparel shopping environmental studies. Moreover, the application of dual coding theory and the

S-O-R paradigm provides valuable theoretical perspectives to understand consumers’ cognitive and behavioral responses in non-store shopping environments.

This study is valuable because appropriate information presentations will contribute to enhancing communications in non-store shopping. Well-presented information can make shopping more efficient for non-store shoppers. The findings of the

compensatory mechanism for haptic information will help non-store retailers understand consumer responses and develop effective marketing strategies using judicious information presentation. By presenting information effectively, non-store retailers can potentially engage consumers in the interactive shopping experience, enhance store image, and ultimately increase competitiveness.

Definition of Terms

The following terms are used in this study.

Attitude toward a product: A consumer’s favorable or unfavorable evaluation of a

product.

Behavioral intention: “A consumer’s judgments about how we will behave in the future”

(Blackwell, Miniard, & Engel, 2001, p.283).

Direct marketing: “An interactive system of marketing that uses one or more advertising

media to affect a measurable response and/or transaction at any location”

(Timmermans & Morganosky, 1999, p.247).

Fabric hand: “Touch feeling of a fabric” (Zeng, Koehl, Sanoun, Bueno, & Renner, 2004,

p.243) and the subjective evaluation of a fabric attained from the sense of touch

(Ellis & Garnsworthy, 1980).

Haptic information: Information obtained through touch by hands (Peck & Childers,

2003a; 2003b).

Imagery: “A mental event involving visualization of concept or relationship” (Luz & Luz,

1978, p. 611) and “a process by which sensory information is represented in

working memory” (MacInnis & Price, 1987, p.433).

Interactivity: the extent to which a user can influence contents of a mediated environment

in a real time (Steuer, 1992) and “interaction with media, a communication

process and a method of controlling a message” (Keng & Lin, 2006, p. 83).

Need for touch (NFT): “A preference for the extraction and utilization of information

obtained through the haptic system” (Peck & Childers, 2003a, p. 431).

Non-store shopping: A way for a consumer to buy without physically visiting a store.

Perceived product quality: A consumer’s evaluations about properties of a product.

Perceived risk: The nature and amount of uncertainty or consequences which consumers

experience regarding the purchase and use of a product (Cox, 1967).

Sensory imagery: “Images that have the qualities of an actual sensation or perception”

(Mazzocco, 2005, p.3). Sensory imagery is accompanied by any sensory quality,

such as visual quality (i.e., visual imagery) and haptic quality (i.e., haptic

imagery).

CHAPTER 2

LITERATURE REVIEW

Overview of Non-store Shopping

The Direct Marketing Association defines direct marketing as “an interactive

system of marketing that uses one or more advertising media to affect a measurable

response and/or transaction at any location” (Timmermans & Morganosky, 1999, p. 247).

Direct marketing strategy has been used by non-store retailers to reach and attract many consumers in their homes. Direct marketing is associated with a variety of formats, such as direct mail catalogs, direct mail, telemarketing, and interactive electronic media (the

Internet) (Blackwell, Miniard, & Engel, 2001). In addition, retailers combine diverse retail operations in order to provide different retail experiences to consumers and to leverage their power of selling. Traditional brick-and-mortar retailers often expand

operations with websites and catalogs, and pure catalogers complete their operations with

an online presence (Pascale, 2000). A survey of multi-channel shoppers showed that 43%

of multi-channel shoppers browsed for products online and bought them in physical retail

stores, 19% of them browsed in catalogs and purchased products in stores, and 16% of

them browsed in a store and purchased online (Schultz, 2004).

In particular, online shopping has emerged as an important format of direct

marketing. A report provided profiles of the 500 largest online retailers which were

ranked by annual sales in 2005 (Brohan, 2006). According to the report of the top 500

online retailers, retail sales for business-to-consumer e-commerce were $109.4 billion with an increase of 25% over 2004’s $87.5 billion. Recently, the merchandise categories of online retailers have been diversified. Although the category of books, CDs, and

DVDs generated the highest online sales in the mid-to late 1990s, various categories of products contributed to overall online retail sales. In 2005, major product categories of the 500 top online retailers consisted of apparel/accessories (22%), specialty non-apparel

(12%), housewares/home furnishings (11%), computer electronics (11%) and sporting

goods (7.8%). Regarding retail sales in each product category, computers and electronics

generated $17.7 billion, office supplies with $10.3 billion, and apparel/accessories with

$7.1 billon (Brohan, 2006). With the growth of online shopping, customer satisfaction with online retailers has been enhanced.

Many of the top 500 online retailers use multi-channel marketing strategies.

Although many retailers are focusing on developing an e-commerce business, catalog retailing still sustains the growth of business and contributes to direct marketing (Brohan,

2006). Traditional catalogers have developed and expanded their online presence so that they reduce costs of paper and postage and provide consumers with a wide range of information (Lohse & Spiller, 1998). Successful catalogers have also become some of the largest online retailers. For example, L.L. Bean Inc. was ranked twentieth among the top

500 online retailers and generated $653.8 million in 2005 (Brohan, 2006).

Characteristics of Online and Catalog Shopping

Benefits of Online and Catalog Shopping

One of the major benefits of online and catalog shopping is convenience.

Consumers are using the Internet and catalogs as a convenient way to save time, costs

and hassles associated with the information search and travel associated with in-store

shopping. A survey reported that consumer satisfaction with online retailers also has

increased by 1.3% from 2004 to 2006 (Guest, 2006). Another recent survey examined the

reasons for consumer dissatisfaction with traditional retail stores and found that consumers were less likely to be satisfied with their purchase experiences at traditional

retail stores than those from catalogs or online because of the relative lack of

merchandise assortment, knowledge, and helpfulness in traditional retail stores

(Consumer Report Buying Guide, 2006). Convenience-oriented shoppers tend to criticize

crowded store conditions, out of stock merchandise, and poorly-trained sales persons

(Seiders, Berry, & Gresham, 2000).

Kaufman-Scarborough and Lindquist (2002) found that convenience is a strong

motive in non-store shopping. Non-store shoppers (i.e., online, catalog, and television

shoppers) prefer non-store shopping since they can shop without leaving their location

and conveniently schedule their time (Kaufman-Scarborough & Lindquist, 2002).

Similarly, Eastlick and Feinberg (1999) found that catalog shoppers shop for clothing and sporting goods to save effort searching, to save time searching, and to shop whenever they wish. Alreck and Settle (2002) found that both catalog and online shopping were viewed as more time saving than traditional shopping and television shopping was

viewed as most time consuming. Catalog and online shopping environments allow consumers to focus on gathering information and making purchase decisions.

Online shopping and catalog shopping provide an economical benefit that is related to money as well as time spent purchasing or conducting product or service information searches (Joines, Sherer, & Scheufele, 2003). Early researchers suggested that online and catalog retailers do business for less (e.g., advertising, marketing and direct distribution of certain goods and information services) compared to brick-and- mortar retailers (Hoffman & Novak, 1996; Korgaonkar, 1984). In particular, online catalogs can save much more in processing costs than traditional catalogs and reduce the time cycle (Hoffman & Novak, 1996). Some non-store retailers provide consumers with relatively low priced products to compete with traditional retailers, and provide an opportunity for shoppers to purchase better quality products for better prices (Eastlick &

Feinberg, 1999).

The availability of information about products and services provide additional benefits to online and catalog consumers and information reduces uncertainty in the purchasing decision process (Kim & Lennon, 2000; Park et al., 2005). This benefit is more evident in online shopping contexts. The Internet and catalogs have the capability to deliver information to any place. In particular, the Internet can contain nearly limitless and extensive information which is available at any time and any place. In addition, the

Internet efficiently and effectively manages, organizes, shares, and disseminates information, and facilitates consumers’ information searching processes (Peterson &

Merino, 2003).

Catalog and online shopping attract consumers by providing enjoyable shopping

environments. Shoppers enjoy exploring online and catalogs to relax, to relieve boredom,

and to obtain product knowledge (Stell & Paden, 1999). Research on catalog and online

shopping segments has identified a recreational segment. Shoppers utilize catalogs and

the Internet for fun and enjoy shopping as a leisure activity. In particular, catalog and

online patronage tends to be more recreational oriented than convenience-oriented.

Recreational shoppers are likely to shop for nonfunctional motives such as fun and

enjoyment. They tend to be attracted by the uniqueness and quality of merchandise and

illustrations and designs of catalogs and websites (Gehrt & Shim, 1998; Kaufman-

Scarborough & Lindquist, 2002; Stell & Paden, 1999). Online shoppers enjoy sharing their information and knowledge through online chat rooms and buyer forums (Kaufman-

Scarborough & Lindquist, 2002). Stell and Paden (1999) also pointed out that recreational catalog shoppers do not engage in shopping in order to solve a particular purchase problem but are willing to leisurely seek information and explore products in catalogs.

Obstacles and Risks of Online and Catalog Shopping

In spite of benefits of online and catalog shopping, consumers perceive a high

level of risk in online and catalog shopping. In both online and catalog shopping,

consumers cannot touch, smell, or feel the products (Vijayasarathy & Jones, 2000). A

survey reported that the biggest obstacle of online apparel shopping is the inability to

directly experience apparel products (48%), followed by security worries (20%), limited

selection (12%), lack of computer or Internet access (11%), expensive delivery/shipping

charges (10%), inconvenience (9%), and discomfort with Internet transactions (8%)

(Transword Business, 2001). Traditional catalog retailers also have difficulties informing and persuading catalog shoppers because of no opportunity for them to directly inspect a product and to evaluate specific features of a product (e.g., color, size, fit) in information search, product evaluation, and purchase situations (Mooy & Robben, 2002). After online and catalog shoppers receive products via mail, they make post-response evaluations while they consume delivered products. They might be dissatisfied with the delivered products in the post evaluation stage. This potential dissatisfaction with products might be exacerbated by the lack of sensory examination in the previous information search stage (McCorkle, 1990). This inability for direct sensory evaluations increases difficulties in evaluation of overall product quality and enhances shoppers’ perceived risk in online and catalog shopping.

Consumers consider whether a product meets performance requirements as expected, which refers to performance risk, and evaluate whether the product has high quality. In online and catalog shopping, uncertainty about product quality and perceived performance risk tend to be high when evaluating clothing because of the lack of direct and sensory evaluations (McCorkle, 1990; Park et al., 2005).

Non-store shoppers also have a concern about the potential loss of status in their social groups (e.g., friends, family) as result of consuming a product, known as social risk, and consider the likelihood of the product having negative effects on consumers’ self- image (i.e., psychological risk). These social and psychological risks tend to be very critical when consumers purchase socially visible products, such as clothing and accessories (McCorkle, 1990). In addition, non-store shoppers may perceive loss of time

when they make a bad purchasing decision and want to return the unsatisfactory product

(Featherman & Pavlou, 2003).

Security is certainly a continuing problem for online shopping. Jones and

Vijayasarathy (1998) examined individuals’ perceptions of online shopping and catalog

shopping and found that individuals have negative perceptions of online shopping security. Consumers are concerned about using credit cards to buy products online and exposing their personal information (Cases, 2002). Both online and catalog shoppers also consider financial loss because of the initial cost of products and potential expenses of repair, maintenance, and exchange (Cases, 2002; McCorkle, 1990).

The lack of access to the Internet is another obstacle in online shopping. Although convenience is a benefit in online shopping, if consumers do not have appropriate hardware and software and a high-speed connection to the Internet, it is impossible to use online shopping (Hoffman, Novak, & Chatterjee, 1995).

In order to increase consumers’ perceived product quality and reduce their perceived risk, online and catalog retailers use diverse risk reduction strategies, such as using endorsements, word-of-mouth advertising, and expensive models; advertising private testing information and major brand images; enhancing brand loyalty; and offering free samples and money-back guarantees (McCorkle, 1990, Roselius, 1971). In particular, apparel shoppers are likely to engage in information seeking activities when they confront perceived risk and a difficulty evaluating apparel quality (Dowling &

Staelin, 1994).

Interactivity

One of the important characteristics of online shopping environments is

interactivity. Interactivity is defined as the extent to which a user can influence contents

of a mediated environment in a real time (Steuer, 1992). Compared to catalog shopping

environments, online shopping environments have a capability of facilitating interaction,

and the interactive nature of the Internet allows consumers to actively engage in a

communication process by controlling a message, amount of information, and order of

presentation at any time (Hoffman & Novak, 1996). Catalog shopping environments tend

to be passive and static in the way they present product information and may reduce

information searching activities of shoppers, whereas online shopping environments

facilitate browsing activities by providing interactive multimedia features, such as chat

functions and advanced technology (e.g., product rotation, zoom functions) (Walsh &

Godfrey, 2000).

Ko et al. (2005) addressed two dimensions of interactivity in the Internet: human- message interaction and human-human interaction. Human-message interaction occurs when Internet users control and customize contents, such as colors, shapes, graphics, sounds, and other message contents. In the Internet environment, users experience interactivity while they are clicking a series of hyperlinks to move to other pages, to select product categories and options, and to explore information presented on websites.

Other examples of interactive activities in the Internet are using keyword search functions, downloading software, playing games and using multimedia features or virtual reality displays on websites. Human-human interaction, the second dimension of interactivity, is associated with reciprocal communication between senders and receivers. In interactive

environments, online retailers can provide product information to consumers, and

consumers can give their feedback on products to marketers, ask about more information about products and service, and request online problem diagnostics. Those activities involve unique features of the Internet, such as email, chat rooms, and real-time feedback

(Ko et al., 2005).

Keng and Lin (2006) defined interactivity as “interaction with media, a

communication process and a method of controlling a message” (p.83) and described

interactivity characteristics in the Internet environments: participant equity characteristics,

dynamic communication process characteristics, mutual understanding characteristics,

and message controlling characteristics. Participant equity characteristics are associated with the ability of the Internet to support exchanges of consumers’ opinions and

communication linkages between consumers. The communication follows dynamic

processes. In online environments, Internet users can exchange and express their opinions

and stop the communication process at anytime. Through mutual communication

processes, Internet users perceive others to be psychologically present and understand

contents of communication. The Internet also allows consumers to control messages by

selecting to choose and add messages (Keng &Lin, 2006). For example, drugstore.com

provides a “customer reviews” service which allows consumers to write their reviews of

beauty products which they have used. Alloy.com provides an opportunity to consumers

to write their opinions of current fashion trends and styles and allows them to email their

friends about product information on the website.

Schlosser (2003) explained object interactivity and navigation interactivity.

Object interactivity refers to users’ direct manipulation of a product in a virtual world.

For instance, using a mouse to press the shutter button of a camera to take a picture and clicking a button to see a closer view of fabrics and alternative views (e.g., side and back views) of a product would be direct manipulation. Navigation interactivity refers to the extent which users move freely through a website in terms of searching, accessing, and retrieving activities. The functions of search engines and hyperlinks facilitate navigation interactivity so that users freely move back and forth through a website and actively explore information about products at deep levels. These functions help shoppers easily compare various products by selecting and controlling contents displayed on websites.

While people manipulate and interact with contents, people may perceive that the mediated contents satisfy their information needs. The increased opportunities to process information by using interactive features may positively influence consumers’ evaluations about a product. Research has studied the effectiveness of interactivity on the

Internet and found effects of interactivity on consumer responses. Ko et al. (2005) examined the antecedents (shopping motivations) and consequences (attitude toward a website, attitude toward a brand, and purchase intentions) of interactivity. The researchers found that consumers who have high information motivations were more likely to engage in human-message interaction on a website and consumers who have high convenience and social interactions were more likely to engage in human-human

interaction on website. In addition, the more consumers engage in human-message and

human-human interactions, the more positively consumers evaluate the website and the

brand (Ko et al., 2005). Keng and Lin (2006) proposed that the extent which people

interact with websites is related to situational and personal involvement. If consumers

have high involvement with shopping goals and high ability to process communication,

they are likely to engage in interactive activities. Keng and Lin found that the high level of interactivity positively influenced recall and recognition for Internet advertising.

Information Presentation

Websites and catalogs are highly information-laden media. Compared to

traditional retail shopping environments, online and catalog shopping enrich the

information environment by providing extensive product and service information.

Consumers engage in search for products and services via the Internet and catalogs in

order to obtain information about them and to compare and evaluate them, and tend to be

attracted by availability of detailed product information (Ward & Lee, 2000). If available

information is not sufficient on a website or in a catalog, consumers may stop to search for information from the website or the catalog and visit other websites and take other catalogs. The importance of information has been emphasized in terms of influencing evaluations of a product, reducing perceived risk, and increasing customer satisfaction and purchase intentions (Kim & Lennon, 2000; Shim, Shin, & Nottingham, 2002). The information environment is based on two fundamental information presentation formats: pictorial and verbal information presentations (Childers & Houston, 1984).

Pictorial Information Presentation

Visual product presentations in retail environments have been emphasized in

relation to visual merchandising. According to McGoldrick (1990), visual merchandising

is a positioning strategy of visible elements and includes two basic elements, visual

displays of products and store design. Research on visual merchandising has emphasized

the important roles of visual merchandising in terms of capturing consumers’ attention, facilitating a process of evaluating products, and influencing consumers’ decisions on product purchases (Lea-Greenwood, 1998, Sen, Block, & Chandran, 2002). Methods of visual product presentation are the most important portions of online apparel visual merchandising (Ha, Kwon, & Lennon, in press). Kim, Kim, and Lennon (2006) pointed out that in online shopping product presentation styles are likely to be more critical when shoppers evaluate apparel as compared to when they evaluate other products, such as books and travel. Interaction between a person’s body and a garment influence fashionability and attractiveness of the garment. How a garment fits one’s body and how colors of the garment match one’s skin tone and hair color can be visually evaluated

(Kim et al., 2006). Manner of product presentation can be effectively used to evaluate these features of clothing in online and catalog shopping.

Ha et al. (in press) examined visual merchandising elements of apparel retail websites and described manner of product presentation (pictorial product information presentation) as one of the visual merchandising elements. Manner of product presentation is defined as a visual merchandising element associated with a way to visually present individual products and product information (Ha et al., in press). Ha et al. divided manner of product presentation into six subcategories: (1) Product view types

(e.g., front view, back view, and side view) and presentation methods (e.g., click-on functions, automatic rotations), (2) detailed views (e.g., 2D larger views, 3D larger views, close-ups, and zoom functions), (3) color and fabric swatch (e.g., fabric swatch, color swatch), (4) product color presentations (e.g., change by color swatch click, colors as

separate items), (5) product display methods (e.g., hanging, mannequin, model), and (6) mix and match.

Li, Daugherty, and Biocca (2003) compared 2D (two-dimensional) product visualization with 3D (three-dimensional) product visualization. The three-dimensional product visualization condition refers to detailed product presentations (e.g., zoom function, close-ups, and rotate function) that facilitate visual, tactile, and behavioral experiences in online shopping environments. On the other hand, two-dimensional product visualization condition does not allow shoppers to interact with the media, such as static product presentation in traditional catalog shopping environments (Li et al.,

2003). According to Ranganathan and Ganapathy (2002), multi-media features, such as animation, rotation, and zoom functions, play an important role in capturing customer attention. Fiore and Jin (2003) also found that click on functions facilitated interactions between visual objects and browsers and influenced attitude toward a product and patronage intentions.

Zoom functions and close-ups allow shoppers to have virtual experiences by clicking on multiple times in order to better evaluate a product (Ha, Kwon, & Lennon, in press). Ha et al. and Kim et al. (2006) also found that many apparel websites provided fabric swatches or color swatches as one method of product information presentation.

Similar to detailed product presentations (e.g., zoom function and close-ups), fabric swatches may help shoppers engage in virtual tactile experiences. Vivid fabric swatches which contain information about the tactile properties of fabric, such as textures, softness, silkiness, and roughness may stimulate virtual experience and tactile simulation.

Compared to online shopping environments, it is impossible to use multi-media

features (e.g., zoom, rotation, click-on functions) in catalog shopping environments.

However, visual appeal has been emphasized in catalog shopping environments. Catalogs

rely heavily on pictures in terms of providing information about a product and need to be

designed to enhance a desirable consumption experience (Mathwick, Malhotra, & Rigdon,

2001). Since catalogs present pictures of products in a limited number of catalog pages,

the manner of visual product presentation might be critical. Some catalogs provide

diverse picture sizes and various views (e.g., front, side, and back views) of an individual

product on high quality paper, whereas some catalogs use white and black pictures. Stell

and Paden (1999) emphasized that effective product presentations in catalogs may

encourage information seeking, enhance vicarious exploration of information and

facilitate imagination of product use.

Verbal Information Presentation

Both catalog and online shopping focus on providing useful verbal information

that facilitates customers’ exploration of products and decision making. Kim et al. (2006)

examined online service attributes which enhance efficient and effective shopping,

purchasing, and delivery. One important service attribute in online apparel websites is

information. Based on Huizingh’s study (2000), Kim et al. identified two categories of

information: commercial information (e.g., company history, general company information) and non-commercial information (e.g., prices, product descriptions, and specifications). Kim et al. maintained that company history and company general

information can be used to increase familiarity with a company and brand name and to

reduce perceived risk. Non-commercial information contains product descriptions, size

charts, shipping costs, sales tax, and so forth. The researchers noticed that apparel

retailers provide different amounts of information. Some websites present basic product

information including color, size, and price information, whereas other websites provide

extensive information such as fabric/fiber information, style and construction details, and

care instructions (Kim et al., 2006).

Traditional stores use labels and hangtags in order to verbally present important

information about products, such as fiber content, size, care instructions. The traditional

shopping context allows consumers to physically touch and handle products as well as

directly examine labels and hangtags of a garment (Hatch & Roberts, 1985). While

shoppers directly and physically examine products, they obtain additional information

through the sense of touch, which is called haptic or tactile information. Typically, the

way people assess garment fabric is by directly feeling the fabric and obtaining haptic

information through touch by hands. For example, physically touching and assessing a sweater provides information about its fabric features, such as softness, weight, and warmth. However, since catalog and online shopping contexts cannot provide the opportunity for shoppers to directly experience products, it is important that they use

information that approximates in-store sensory or experiential information (Park & Stoel,

2002).

Fabric hand descriptions might be essential for efficient communication of haptic

(tactile) sensory specifications between consumers and online and catalog retailers

(Soufflet, Calonnier, & Dacremont, 2004). Fabric hand is defined as the “touch feeling of

a fabric” (Zeng, Koehl, Sanoun, Bueno, & Renner, 2004, p.243) and the subjective

evaluation of a fabric attained from the sense of touch (Ellis & Garnsworthy, 1980). In

general, fabric hand is assessed in two ways: (1) subjective evaluation and (2) objective

evaluation. The subjective assessment of fabric hand is based on the evaluation of experts

who touch a fabric with their hands and the objective evaluation is done by instrumental

measurements (e.g., Kawabata Evaluation System for Fabrics) (Zeng et al., 2004). Fabric

hand descriptions include information about a fabric’s roughness, softness, smoothness,

thickness, weight, firmness, elasticity, stiffness, warmth and so forth. These properties of

a fabric are evaluated when the fabric is being touched (Soufflet et al., 2004; Zeng et al.,

2004). In non-store shopping contexts, detailed fabric hand descriptions may help

shoppers indirectly obtain information about haptic properties of a fabric. Soufflet et al.

maintained the need for accurate and appropriate fabric hand descriptions which can

stimulate richness of individual’s sensory perceptions. Soufflet also described three haptic dimensions utilized to perform perceptions of fabric hand (i.e., haptic perceptions).

The three dimensions are soft vs. harsh dimension, the thin vs. thick dimension, and the

supple vs. stiff dimension. Among the three dimensions, softness contains high communicative value to stimulate haptic perceptions (Soufflet et al., 2004).

This current study focuses on investigating the roles of pictorial and verbal

information in influencing consumer responses and on examining how customers respond

differently to different shopping environments (catalog and online shopping

environments). The critical problem with assessing products in catalog and online apparel

shopping environments is the inability for direct examination of products. Certain

pictorial (e.g., fabric swatch) and verbal information (e.g., fabric hand descriptions) may stimulate perceptions of haptic information and simulate how a garment and a fabric

feels. The next section presents theoretical frameworks and hypothesized model of the

study.

Theoretical Frameworks

Dual Coding Theory and Imagery

Dual Coding Theory

Dual coding theory provides a framework of this study. Dual coding theory

explains the independence and interconnectedness between pictorial and verbal

information and the effects of information evoking imagery on cognition (Paivio, 1971;

1975). The basic assumption of the theory is that cognitive activity is handled by two

independent, but partially interconnected subsystems: One system is an image (imagery)

system which specializes in representation and processing of information regarding

nonverbal objects and events and in evoking mental images. The other system, the verbal

system, deals with processing verbal (linguistic) information. The independence of the two systems means that verbal and image systems are each independently activated without the other. For example, we can remember an event with nonverbal images and without verbal descriptions. At the same time, since the two systems are interconnected, one system can activate the other system. For example, verbal descriptions stimulate our memory regarding images of an event and a situation (Paivio, 1986; Paivio & Desrochers,

1980).

Nonverbal information, such as a picture or an object, is processed and organized in a synchronous or simultaneous manner. For example, parts of a face (e.g., eyes, nose,

and lips) are simultaneously seen and synchronously illustrated by a mental image of the

face. The face can be a part of a larger image (e.g., a human body) as well. Nonverbal

(pictorial) information tends to be more easily and rapidly encoded and switched among

parts and wholes as compared to verbal information. On the other hand, verbal

information is sequentially processed in the way that smaller units are organized into

larger units. Syllables are organized into words, words into phrases, and so on, up to

entire written works. The simultaneous processing of pictorial information tends to be

facilitated more quickly than the sequential processing of verbal information (Paivio,

1986; Paivio & Desrochers, 1980). Pictorial information is directly encoded by using both images and words, whereas verbal information is encoded mainly verbally.

Although concrete verbal information tends to evoke mental images, the imagery evoked by verbal information is indirect as compared to direct as in the case of pictures. Thus, the dual coding of images makes them more easily accessible in long-term memory and results in picture superiority effects (Paivio, 1986; Childers & Houston, 1984).

Picture Superiority Effects and Imagery

The picture superiority effect refers to the fact that images (pictures) are more memorable than words. Greater memory for pictures over words (the picture-superiority effect) can be explained by the mental imagery process. Imagery is defined as “a mental event involving visualization of a concept or relationship” (Lutz & Lutz, 1978, p. 611) and “a process by which sensory information is represented in working memory”

(MacInnis & Price, 1987, p. 473). The picture-superiority effect on memory is explained by visual imagery (Childers & Houston, 1984). Since visual imagery stimulates cognitive

elaboration, it helps to create a strong memory trace, enhances learning and retention of

material, and increases the likelihood that information will be retrieved (Paivio & Foth,

1970).

The effectiveness of advertisements has been measured by recall and recognition.

After people are exposed to pictures or verbal descriptions, they are asked to describe an

advertisement, a product or a brand (recall), and to recognize whether they have seen a

specific description or a picture (Blackwell et al., 2001). Luz and Luz (1977) studied the

effect of interactive pictures, noninteractive pictures, and their verbal descriptions on

recall and recognition. Interactive pictures described brand name and product class in a

visual format, noninteractive pictures showed brand name or product class in a visual

format, and the control condition consisted of only verbal descriptions without a picture.

In that study, interactive pictures were superior to words in terms of recall and learning

features of a product and a brand. Childers and Houston (1984) studied the effect of

pictorial and verbal information on immediate and delayed recall, and found that pictures

were more quickly recalled than verbal information when memory was measured over

time. Rossiter and Percy (1978) found that visually oriented advertising (a large picture

of a product with small verbal copy underneath) had a significant impact on positive

attitude toward a brand as compared to verbally oriented advertising (large verbal copy

with a small picture of the product underneath). In general, pictures are considered to be a

relatively vivid type of information and to be more attention-getting and easier to process

as compared to words.

Vividness Effects

Imagery processing is facilitated by vivid and concrete sensory representations of feelings and memories. Vividness is defined as the clarity of images (McInnis, & Price,

1987). According to Edell and Staelin (1983), pictorial information in print

advertisements has a critical impact on memorability of advertisements and cognitive

evaluations of products after people are exposed to the advertisements. An imagery

process engages a single sensory dimension or multiple sensory dimensions (e.g., smell,

taste, sight, or tactile sensations) (Maclnnis & Price, 1987; Yuille & Catchpole, 1977).

High imagery-provoking pictures stimulate haptic imagery or tactile perceptions, such as

roughness of an object (Cairns & Coll, 1977; Klatzky, Lederman, & Matula, 1993). As

compared to abstract pictures, vivid and concrete pictures are associated with objects or

things that can be seen, heard, smelled, or tasted (Rossiter, 1982). The lack of sensory

dimensions of information in working memory makes information less concrete and more

abstract. Concrete and vivid pictorial information presentations tend to facilitate visual imagery and cognitive elaboration of stimulus-relevant information in memory, and finally enhance consumers’ judgments about a purchase (Kisielius & Sternthal, 1984;

MacInnis & Price, 1987). Babin and Burns (1997) found that a concrete picture in a print advertisement influenced imagery processing more positively and resulted in more favorable attitude toward a brand and an advertisement than a less concrete picture or no picture.

Concreteness and vividness of verbal information may activate images, help us visualize, and refer to tangible objects (Hong, Thong, & Tam, 2004). In the study by

Rossiter and Percy (1978), concrete copy was more effective than abstract copy. Some

words, called high imagery words, can evoke images in individuals’ minds (Unnava &

Burnkrant, 1991). The effect of high imagery words can also be explained by dual coding

theory. For example, high imagery words describe textures and features of products and

how they feel to the touch, and low imagery words describe factual information about a

product, such as price and fiber content (Fiore & Yu, 2001). High imagery words help

people evoke images in their minds, understand the message, and are superior for

memory retrieval as compared to low imagery words (Unnava & Burnkrant, 1991).

Sensory Imagery

Although images are usually referred to as visual mental representation, they

could refer to tactile, auditory, olfactory and gustatory mental ideas and images according

to the sense modality involved. Sensory imagery is associated with the qualities of an

actual sensation or perception. Concrete images convey vivid sensory qualities such as

visual, auditory, tactile, gustatory, or kinesthetic (Roeckelein, 2004). Images that

combine more than one sensory modality are called composite images (Drever,

1952/1973).

Reber (1995) points out that a picture is not restricted to evoking visual imagery, but can evoke tactile, auditory, gustatory, or olfactory images. Paivio (1986)’s dual coding theory is also based on the nature of symbolic systems (i.e., verbal and nonverbal

systems). The symbolic representations contain features of different sensory and response

modalities (i.e., visual, auditory, haptic, taste, smell systems). For example, visual

sensory modality is associated with verbal and nonverbal encodings (symbolic systems).

Visual verbal encodings can be a mental representation of a visual word form, such as

‘dress.’ A nonverbal visual encoding can be a mental representation of a visual object or scene, such as a summer dress. The nonverbal (imagery) system consists of a set of interconnected parts specialized for dealing with sensory systems. For example, a person may visualize a taffeta dress with or without imagining the rustling sound of the dress or imagining how it feels when it is touched. One sensory system can be functionally mapped onto any other. In particular, the visual system plays a dominant role in the transfer from one sensory system (e.g., sight) to the other (e.g., touch). The verbal system is also associated with separate subsystems (e.g., speaking, reading, and writing) and the subsystems are interconnected with each other (Paivio, 1986).

Most research on sensory information processing has studied visual imagery, whereas few studies have studied other sensory imageries. In particular, there is a lack of research on haptic imagery. Sensations of touch involve the contact of objects to movement of the skin of a body part and active exploration of the object’s properties by using skin, muscles, and joints (Gibson, 1962). Perceptions of touch include texture, weight, temperature, size and shape and the haptic system is activated to encode those material properties of an object (Klatzky, Lederman, & Matula, 1991). Haptic information is defined as information attained through touch by hands (Peck & Childers,

2003a; 2003b).

Researchers in psychology have discussed visual-tactile interactions (Igarashi,

Kitagawa, & Ichihara,2004; Klatzky, Lederman, & Matula, 1991; 1993). Interactions between vision and touch facilitate access to information about haptically accessible object properties (Klatzky, Lederman, & Matula, 1991). For example, when we think about the smoothness of a pane of glass, we can imagine haptic and visual sensations by

mentally observing our hand touching the glass (Katz, 1989). People evaluate spatial images by simultaneously using vision and touch (Kerr, 1983). Since haptic images and visual images are simultaneous, the interdependency between haptic and visual imagery may be explained (Katz, 1989).

Vision plays an important role in perceiving haptic information (Igarashi et al.,

2004). If we see our own bodies in a mirror, we can receive haptic information and tactile performance through mirrored images. The mirrored images reflect interrelationships between our three-dimensional bodies and clothing. The visual images help us perceive textures of our garments (e.g., texture, softness, drapability). When only visual exploration (e.g., pictures) of an object are available and touch exploration is absent (e.g., catalog and Internet shopping contexts), can visual information provide haptic information and influence the interactions between vision and touch? When we browse pictures or play a game through a computer screen, we manipulate a computer mouse.

Manipulated visual objects convey visual information and may provide tactile information. Once we are accustomed to manipulaing visual information, we may sometimes receive touch sensations of details of pictures from the computer screen

(Igarashi et al., 2004). However, research reported that if touch is only available without vision, judgments of precise shape information tend to be inaccurate as compared to when both vision and touch are allowed or when vision is only available (Bryant & Raz,

1975).

In non-store shopping contexts, pictorial information and verbal information may serve to compensate for the inability to touch a product by stimulating retrieval of information about a product’s expected haptic properties based on information stored in

memory (Peck & Childers, 2003a). For example, pictures and verbal descriptions which focus on providing haptic information (e.g., texture) might play a role in facilitating haptic imagery and helping to imagine and retrieve memories.

Stimulus-Organism-Response (S-O-R) Paradigm

Along with dual coding theory, this study is based on the Stimulus-Organism-

Response (S-O-R) paradigm. Mehrabian and Russell (1974) produced the S-O-R

paradigm for explaining relationships between the physical environment and human

behavior. The paradigm explains that environmental stimuli (S) cause consumers’

evaluations (O) and the evaluations prompt them to approach or avoid (R) the

environment. Research in environmental psychology has applied Mehrabian and

Russell’s S-O-R paradigm to investigate in-store shopping environments (Spies, Hesse &

Loesch, 1997; Turley & Milliman, 2000). Baker, Parasuraman, Grewal, and Voss (2002)

classified multiple store atmospherics into three categories: Social factors (e.g., employee

perceptions), design factors (e.g., layout, style) and ambient factors (e.g., music, scent).

Turley and Milliman (2000) classified five dimensions of atmospherics: External

variables (e.g., exterior signs, color of building), interior variables (e.g., flooring,

lighting), store layout (e.g., floor space, traffic flow), interior displays (e.g., product

display), and human variables (e.g., employee and customer characteristics).

Recent research has emphasized that non-store shopping provides unique

attributes of environmental cues which are different from traditional shopping

environments. Both catalog and Internet shopping lack environmental characteristics of

traditional stores, such as social cues, tactical and olfactory cues, and are more likely to

focus on pictorial and verbal cues (Eroglu, Machleit, & Davis, 2003). Internet shopping

contexts provide certain interactivity functions, such as e-mail inquiries to sales associates, voice and video applications, and automated price comparisons (Vijayasarathy

& Jones, 2000). According to Eroglu, Machleit, and Davis’s model (2001), online

environmental stimuli (S) are divided into two categories: High task relevant cues and

low task relevant cues. High task relevant cues are related to shopping goals (e.g.,

descriptions of products, price, and sales information), whereas low task relevant cues are

not relevant to the shopping task (e.g., background colors, music, and decorative pictures).

Online environmental cues influence affective (e.g., emotions) and cognitive internal states (e.g., attitudes, beliefs, attentions), which intervene between online environmental cues and approach or avoidance responses (e.g., intentions to stay, spend, and return)

(Eroglu et al., 2001) (see Figure 2.1). Eroglu et al. (2003) found that online environmental cues (high/low task relevant information) significantly influenced consumers’ emotions, along with attitudes, satisfaction, and approach/avoidance behaviors. They also found that individual level of atmospheric responsiveness and involvement moderated relationships between online environmental cues and consumers’ affective (emotions) and cognitive states (attitudes).

Stimulus Organism Responses Involvement Online Environmental Cues Internal States Shopping Outcomes

High Task Relevant Affect Approach Low Task Relevant Atmospheric Cognition Avoidance Responsiveness

Figure 2.1 An S-O-R model of consumer response to online shopping (Eroglu, Machleit,

& Davis, 2001, p. 179)

Researchers consistently have recognized the importance of store environmental

cues. Store environments are considered significant marketing tools for retailers to differentiate themselves from competitors and to enhance store image (McGoldrick &

Pieros, 1998). It is important that store environmental elements should be created and

designed based on customers’ responses. Direct experience (e.g., trying on products,

feeling textures of fabrics) plays an important role in purchase decisions of apparel

shoppers (Eckman, Damhorst, & Kadolph, 1990). In traditional apparel shopping

contexts, the store environment is designed to focus on enhancing direct consumer

experiences by creating effective product displays (e.g., shelf space, special displays) and

store layouts (e.g., traffic flow, fitting rooms) and providing social cues (e.g.,

number/friendliness of employees). However, apparel catalog and online environments

are limited in the tangibility of product information that can be provided and try to reduce

these limitations by providing pictorial information (e.g., pictures, larger view) and

verbal information (e.g., product descriptions). Therefore, pictorial and verbal

information presentation is important in the non-store shopping context.

The process of mental imagery is applicable to the S-O-R paradigm. Organism

(O) must be associated with stimuli (S) and response (R). The relationships between S and O would entail the effect of any stimuli on mental representations, and the

association between O and R would explain the effect of the mental operation on customer responses. The association between S and R does not need to be learned and O

needs to vary according to different stimuli (Kosslyn, 1980). It seems that imagery satisfies theses associations. Information gives rise to mental images (S-O link) and we

react to our mental images (O-R link). Mental images can differ according to different

information presented (Mazzocco, 2005).

Hypothesis Development

Based on the S-O-R paradigm and the dual coding theory, this study developed a

new model to examine how verbal and pictorial information presentation influence

consumer internal states and responses in the contexts of catalog and Internet apparel shopping. The overall sequence of effects in the model of the study is that pictorial and

verbal information (environmental cues) influence consumers’ perceived product quality

and perceived risk (consumers’ internal states). The dual coding theory explains the

effects of diverse pictorial and verbal information formats on consumers’ internal states.

Consumers’ individual characteristics (i.e., NFT) moderate the relationships between

information presentations (pictorial and verbal information) and consumers’ internal

states. Then, consumers’ internal states influence behavioral intentions (shopping

outcomes) (see Figure 2.2).

Stimulus Organism Response (Environmental Cues) (Internal States) (Shopping Outcomes)

Information Presentation Perceived Product Quality Pictorial information

Attitude Behavioral toward Intentions Verbal Information a Product

Shopping context Perceived Risks

Individual characteristics

NFT

Figure 2.2. Research model: The effects of information on customer internal states and shopping outcomes in non-store apparel shopping contexts

The Effect of Environmental Cues on Internal States

In non-store shopping contexts, pictorial and verbal information play important

roles in influencing consumers’ internal states. In the S-O-R paradigm, internal states refer to consumers’ affective and cognitive states including mood (Eroglu et al., 2001;

2003; Spies et al., 1997) and attitudes (Eroglu et al., 2001; 2003), perceived product quality (Davis, 1985, Eckman et al., 1990), perceived risk (Koering, 2003; Park et al.,

2005), and store image (Schlosser, 1998). Perceived product quality, perceived risk, and attitudes toward a product are considered to be internal states in the current study.

Perceived product quality

Consumer psychologists have studied the factors which influence consumers’

perceptions of product quality and commonly divide product evaluative cues into two

classes: intrinsic and extrinsic cues (Jacoby, Olson, & Haddock, 1971; Olson & Jacoby,

1972). Intrinsic cues refer to the product’s inherent characteristics that cannot be

manipulated without altering the physical attributes of the product itself, such as design

or style, whereas extrinsic cues are defined as non-physical product properties that can be

changed without altering the functional nature of the product, such as price, brand name

or store name (Eckman et al., 1990; Olson & Jacoby, 1971). Many studies of perceptions

of product quality have used a single-item scale (Valenzi & Andrews, 1971; White &

Cundiff, 1973). Several researchers have tried to use and develop multi-item measures of

perceived product quality. Davis (1985) used a scale measuring perceptions of apparel

quality in terms of construction quality, fabric quality, quality of the notions, quality of

design, overall quality, fashionability, status and uniqueness. Eckman et al. (1990)

identified four categories for assessing customer perceptions of apparel quality: (1) aesthetic criteria (e.g., color/pattern, styling, fabric, and appearance), (2) usefulness criteria relating to utilitarian concerns (e.g., versatility, matching, and appropriateness),

(3) performance and quality criteria (e.g., fit, comfort, care and workmanship), and (4) extrinsic criteria (e.g., price, brand, and competition).

Stone-Romero, Stone, and Grewal (1997) developed a measure of perceived product quality (PPQM) which included four dimensions: flawlessness, durability, appearance, and distinctiveness. The flawlessness dimension of perceived product quality refers to consumers’ beliefs about the number and types of defects in a product and involves the production of a product based on standards of shape, size, fit, strength and so on. Durability is defined as individuals’ beliefs about the life-expectancy of a product

(Garvin, 1987). The physical appearance of products is the third dimension of the measurement of product quality. Examples are evaluations of designs and styles of a garment. The last dimension of PPQM is distinctiveness dealing with extrinsic cues.

Distinctiveness and uniqueness enhance the status of a product and the evaluations of quality of a product. Perceptions of product quality are improved by distinctiveness of a product regarding increased price of a product, positive store image, and well-known brand names (Stone-Romero, Stone, and Grewal, 1997). Forsythe (1991) examined the effect of intrinsic and extrinsic cues on evaluations of apparel quality and found intrinsic cues (actual garment characteristics) are more important factors than extrinsic cues

(brand names) in evaluating garment quality. On the other hand, Wheatley, Walton, and

Chiu (1977) found that perceived product quality was enhanced by high prices and well- known brand names.

Perceived Risk

Perceived risk is defined as the nature and amount of uncertainty or consequences which consumers experience regarding the purchase and use of a product (Cox, 1967).

Cunningham (1967) identified six categories of perceived risk: Performance, financial, opportunity/ time, safety, social, and psychological loss. Simpson and Lakner (1993) examined perceived risk in catalog apparel shopping and found four components: social/psychological risk (e.g., fashion innovativeness and acceptance, and conformation to others), economic risk (e.g., loss of money from purchase clothing), performance risk

(e.g., loss associated with style and lack of durability), and physical risk (e.g., bodily discomfort, appearance). Forsythe and Shi (2003) explored four components of perceived risk in Internet shopping: financial, product performance, psychological, and time/convenience loss risk.

Researchers have discussed the important roles of effective information presentation. The availability of information influences information search, perceptions of uncertainty and risk, purchase intentions, and purchase decisions (Cox, 1967; Kim &

Lennon, 2000). The information about a product (e.g., price, color, fiber content, care instructions, country of origin) can be utilized to evaluate a product and make purchase decisions (Cox, 1967). Kim and Lennon (2000) found that people who perceived more information perceived less risk and had greater purchase intention as compared to people who perceived less information in television shopping contexts. Research in catalog apparel shopping found that non catalog shoppers tend to perceive higher risk than catalog shoppers and suggested that product information, such as detailed product

descriptions, may play a role in reducing perceived risk (Kwon, Paek, & Arzeni, 1991;

Simpson & Lakner, 1993).

Online shoppers are also attracted by the availability of detailed product

information and the variety of choices offered (Ward & Lee, 2000). Grewal, Munger,

Iyer, and Levy (2003) maintained that value-enhancing information, such as online-

security guarantees and money-back guarantees, can reduce perceived risk and increase

online shoppers’ price expectations and their willingness to pay more. Li, Tan and Xie

(2002) emphasized that web-based service quality is evaluated based on quality of

information. Cases (2002) maintained that pictorial information of products contributed

to reducing risk and that website designers should try to convey product quality by using new innovations (e.g., zooming in on image of products).

The features of pictorial and verbal components influence consumers’ evaluations about a product. Holbrook and Moore (1981) found that pictorial displays had more positive impacts on customer judgments than verbal descriptions. Since pictures tend to get more attention, be pleasant, and be easily processed (Edell & Staelin, 1983), pictures facilitate more favorable attitudes toward a product and purchase intentions than verbal information does (Mitchell & Olson, 1981; Peck & Childers, 2003a; Percy & Rossiter,

1983).

In non-store shopping contexts, combinations of pictorial and verbal information may have an impact on consumers’ internal states positively or negatively via imagery

(Peck & Childers, 2003a; 2003b). When consumers are unable to directly examine a product in the catalog or Internet shopping context, high imagery information may play an important role in stimulating retrieval of haptic information about a product (e.g.,

texture, weight) which is stored in memory (Peck & Childers, 2003a). High imagery

information compensates for the lack of physical contact, substitutes for consumption

experiences, attracts non-store shoppers to explore websites or catalogs, and leads to

favorable product evaluations (MacInnis & Price, 1987; Mckinney, Yoon, & Zahedi,

2002). The compensatory effect of pictorial information and verbal information for haptic

information has been discussed. Pictorial information may be more likely to compensate

for haptic information than verbal information. Peck and Childers (2003a) studied how

pictorial and verbal information compensate for the lack of haptic information. Peck and

Childers found that pictorial and verbal information containing high haptic imagery (e.g., cellular telephone weight and sweater softness) tended to reduce frustration associated with product evaluations and positively influenced perceptions of product quality. Fiore and Yu (2001) found that imagery copy (i.e., text) and fabric samples positively influenced pre-purchase approach responses and attitudes toward a product in a catalog apparel shopping context. Based on this rationale the following hypotheses were developed.

H1: Pictorial information associated with high haptic imagery will have a positive effect on consumer internal states (a) more positive perceptions of product quality and b) less perceived risk) compared to pictorial information associated with low haptic imagery.

H2: Verbal information associated with high haptic imagery will have a positive effect on consumer internal states (a) more positive perceptions of product quality and b) less perceived risk) compared to verbal information associated with low haptic imagery.

H3: Pictorial information and verbal information will interact to affect consumer internal states (a) perceived product quality and b) perceived risk).

Situational Differences

Situational characteristics have been considered important in information processing. In-store shopping environments provide consumers with the opportunity for direct product experience, whereas non-store shopping (e.g. Internet shopping and catalog shopping) is associated with indirect experience (Burke, 1997; McCabe, &

Nowlis, 2003; Mooy & Robben, 2002). Direct product experience occurs when a

consumer directly interacts with a product using full sensory capacity (visual, auditory,

taste, smell, and haptic) (Gibson, 1966). For example, apparel shoppers can examine

texture, color, and weight of an apparel product and how it fits their bodies in the in-store

environment (Eckman et al., 1990; Fiore & Jin, 2003). Direct product experience, which

is normally gained from in-store shopping environments, tends to enhance the ability to process product-related information and is more likely to increase consumers’ positive attitudinal confidence than indirect experience (Mooy & Robben, 2002).

Indirect experience occurs when a consumer cannot have direct sensory contact (e.g., via pictures, product descriptions, or perusing a product in store window) and cannot fully interact with a product (Mooy & Robben, 2002). Indirect experience is mostly obtained

in the non-store shopping context. It is difficult for non-store retailers to simulate haptic

information (Biocca, Kim & Levy, 1995). Catalog retailers try to provide haptic

information to enhance positive consumer responses. Some catalog retailers provide

samples of a product (e.g., beauty product samples or fabric samples) (Fiore & Yu, 2001).

Online retailers also try to reduce the inability to directly examine products and enhance the ability to select and control the form and content of an environment in real time, which is called interactivity (Steuer, 1992; see Li, Daigherty, & Biocca, 2002). Compared to the catalog shopping context, in the Internet shopping context advanced technology

(e.g., product rotation, zoom function) may facilitate interactivity and provide visual sensory information. Research on Internet shopping environments has found that high interactive advertising (e.g., move, rotate, zoom in or out, mix and match) can be associated with more active cognitive and affective activities, more favorable attitudes toward a product, and greater purchase intentions than low interactive advertising (e.g., indirect experience, static information presentation) (Fiore & Jin, 2003; Li, Daugherty, &

Biocca, 2001). For example, both Land’s End.com and Lanebryant.com provide a 3D simulation function (My Virtual Model). Shoppers can create a virtual model by selecting body sizes, hair color and style, facial shapes, and try products on the model. Other online apparel retailers also provide larger views (zoom in or out), different angles of products (side or back), and other rotation functions. Because the Internet context is interactive the following hypothesis was developed.

H4: The Internet shopping context will have a more positive effect on internal states (a) more positive perceptions of product quality and b) less perceived risk) as compared to a catalog shopping context.

The Moderating Effect of Individual Characteristics

Need for Touch (NFT)

Research in marketing and consumer behavior examined individual differences in

consumer imagery information processing. Most research has studied individual

differences in pictorial and verbal information processing (Childers, Houston & Heckler,

1985). Peck and Childers (2003a) emphasized differences in customer preference or

motivation for touch (need for touch). Need for touch (NFT) is defined as “a preference

for the extraction and utilization of information obtained through the haptic system”

(Peck & Childers, 2003a, p. 431).

According to Peck and Childers (2003b), people who are more haptically oriented

(high in NFT) are more likely to chronically access haptic information and to attend to

haptic information more than people who are less haptically oriented (low in NFT). On

the other hand, low NFT consumers are more likely to be satisfied with pictorial

information even though they would prefer to evaluate the texture of a product. Peck and

Childers found that when a picture (of a cellular phone or a sweater) was provided, low

NFT people were more confident in their judgments and had more positive perceptions of quality whether or not a written haptic description was provided. Peck and Childers explained that for the low NFT people, the picture may suffice in the retrieval of information from memory about the product’s properties. When a picture was provided (a cellular phone) with a written haptic description, high NFT people were more confident

in their judgments and had more positive perceptions of product quality compared to when no haptic description and no picture were presented. The haptic description might play a role similar to high imagery words in helping to imagine and retrieve memories.

Although there is little research on the moderating effect of NFT, it is expected that NFT may moderate the effect of information presentation on consumers’ perceived product quality and perceived risk. This discussion led to the development of the following hypotheses.

H5: Need for touch will moderate the relationship between information presentation and consumer internal states (a) perceived product quality and b) perceived risk).

The Relationships between Internal States and Consumer Responses

Attitude is defined as “a person’s favorable or unfavorable evaluation of an object” (Fishbein, & Ajzen, 1975, p. 12). Overall attitudes toward a product depend on how consumers positively or negatively perceive product quality and risk. Teo and

Yeong (2003) maintained that overall deal evaluation in Internet shopping is positively associated with perceived product quality and negatively associated with perceived risk.

Thus, the study expects a positive relationship between perceived product quality and attitudes toward a product and a negative relationship between perceived risk and attitudes toward a product.

Based on the S-O-R paradigm, the effects of internal states (perceived product quality, perceived risk and attitudes toward a product) on shopping outcomes (behavioral intentions) are also expected. Purchase intention refers to an intention of a consumer to purchase a product in a shopping situation. Based on the theory of reasoned action

(Fishbein & Ajzen, 1975), behavioral intentions are a function of two determinants:

attitude toward the object or behavior and subjective norm, which is an individual’s perception of normative social pressure to perform or not perform the particular behavior.

Past research has found relationships among perceived product quality, perceived risk, attitudes toward a product, and behavioral intentions. Shim et al. (2001) and Yoh,

Damhorst, Sapp, and Laczniak (2003) found attitudes toward Internet shopping were positively related to apparel buying intention through the Internet. When consumers have positive perceptions of product quality, perceive low risks regarding purchase or use of a product, have favorable attitudes toward a retailer, and perceive high satisfaction with their purchases, they are likely to purchase the product, to engage in positive word-of- mouth-communications, and to pay extra money for the product (Athanassopoulos,

Gounaris, & Stathakopoulous, 2001; Reichheld & Sasser, 1990; Park et al., 2005).

However, if customers perceive a high amount of risk from a purchase, have negative perceptions of product quality, and have negative attitudes toward a retailer and experiences in purchasing from a store, they tend to engage in complaining and switching behavioral responses (e.g., switching to another retailer, complaining to external agencies, and doing less business with the company) (Athanassopoulos et al., 2001; Dick & Basu,

1994; Vijayasarathy & Jones, 2000; Zeithamal, Berry, & Parasuraman, 1996). Therefore, the following hypotheses were developed.

H6: Perceived product quality will be positively associated with attitude toward a product.

H7: Perceived risk will be negatively associated with attitude toward a product.

H8: Perceived product quality will be positively associated with behavioral intentions.

H9: Perceived risk will be negatively associated with behavioral intentions.

H10: Attitude toward a product will positively influence behavioral intentions.

CHAPTER 3

PRETESTS

The proposed model and hypotheses were examined in an experimental study.

The experiment was conducted (1) to examine the effects of pictorial and verbal information on consumer internal states (i.e., perceived risk, perceived product quality,

and attitude toward a product), (2) to investigate the effects of shopping contexts on

consumer internal states, and (3) to examine the moderating role of need for touch (NFT)

on the process by which pictorial and verbal information influences consumer internal

states. The experimental design of this study was a 2 (pictorial information: picture swatch or no swatch) x 2 (verbal information: high haptic imagery description or low haptic imagery description) x 2 (shopping context: online shopping or catalog shopping) x 2 (need for touch (NFT): high NFT or low NFT) between-subjects factorial design.

This chapter explains pretests conducted to select and develop stimuli for the main experiment in this study. This phase of the chapter consists of (1) stimulus selection of one apparel category (pretest 1), (2) stimulus selection of two apparel items (pretest 2),

(3) stimulus development of pictorial information (pretest 3), (4) content analysis (pretest

4), and (5) stimulus development of verbal descriptions (pretest 5) (See Figure 3.1).

Step 1: Stimulus Selection and Pretest (Chapter 3)

Pretest 1: Browse apparel categories and select one appropriate product category for the study

Pretest 2: Evaluate ten different examples of one apparel category to select two stimuli for the main experiments

Pretest 3: Evaluate pictorial information (picture swatch and no picture swatch) in terms of vividness and haptic imagery elaboration

Pretest 4: Content analysis for extracting verbal descriptions of dresses

Pretest 5: Evaluate verbal descriptions of dresses and develop two sets of high and low haptic imagery descriptions. Test the differences between high and low haptic imagery differences in haptic imagery-evoking ability

Step 2: Main Study (Chapter 4)

Browse a mock website or a mock catalog and select one apparel item. Then, answer a questionnaire.

Figure 3.1. The procedure of the study

Pretest 1

The purpose of the first pretest was to select one appropriate product category for

use in the main experiment. One apparel category which is more likely to be touched and

tried on by shoppers should be chosen since this study focuses on the effect of haptic

information about a product. A convenience sample of female college students who were

enrolled in an undergraduate class in the Textiles and Clothing program at the Ohio State

University voluntarily participated in Pretest 1. Sixty female students evaluated 8 product

categories (e.g., blazer, sweater, dress) in terms of the importance of touch and fit of the

clothing items in the decision to purchase them using 7-point Likert-type scales (1: not

extremely important and 7: extremely important) (See Appendix B). Dresses were

selected because they received higher scores on the importance of both touch and fit as

compared to other clothing categories, such as sweater and pants (See Table 3.1). A

repeated measures MANOVA was conducted with the eight apparel categories as a

within-subject factor and importance of touch and importance of fit as dependent

measures, and revealed a significant difference among the apparel categories for

dependent measures (Hotelling’s T2 = 2.28, F (14, 45) = 7.31, p = .00) (See Table 3.2).

Paired comparison t-tests were conducted to compare dependent measures between

dresses and other apparel categories. The results of paired comparison t-tests revealed

that touch of dress was significantly more important than touch of blazer, coat and skirt.

And for the importance of fit, dress was significantly more positively rated than shirt, sweater, coat, and top (See Table 3.3).

Mean S.D. Min. Max. Importance of touch Sweater 5.98 1.00 3 7 Dress* 5.77 0.98 3 7 Shirt 5.72 1.08 2 7 Top 5.67 1.20 3 7 Pants 5.58 1.16 3 7 Skirt 5.20 1.05 3 7 Coat 5.17 1.17 3 7 Blazer 5.12 1.11 2 7 Importance of fit Pants 6.83 0.46 5 7 Dress* 6.72 0.64 4 7 Blazer 6.72 0.52 5 7 Skirt 6.53 0.68 5 7 Shirt 6.30 0.91 3 7 Coat 6.28 0.85 4 7 Top 6.18 1.02 3 7 Sweater 6.05 0.89 4 7 Note. * Items selected for the main study.

Table 3.1. Ratings of Apparel Categories

MANOVA Result ANOVA Result

Hotelling’s F Partial F Partial Source Dependent variables p p T2 (14, 45) η2 (4, 58) η2 Apparel Importance of touch 2.28 7.31 .000 .695 3329.47 .000 .983 category Importance of fit 9982.98 .000 .994

Table 3.2. Repeated measures MANOVA and ANOVA Results for Importance of Touch and Fit by Apparel Categories

Paired Difference Paired t-Test Result Dependent Paired mean measures comparison M SD SE t df p Importance of touch Dress – Blazer .65 1.16 .15 4.33 59 .000 Dress – Shirt .05 1.17 .15 .33 59 .742 Dress – Sweater -.22 1.04 .13 -1.61 59 .113 Dress – Coat .60 .98 .13 4.75 59 .000 Dress – Skirt .57 1.11 .14 3.95 59 .000 Dress – Pants .17 1.30 .17 1.00 59 .321 Dress – Top .10 1.11 .14 .69 59 .490 Importance of fit Dress – Blazer .00 .78 .10 .00 59 1.000 Dress – Shirt .42 .79 .10 4.10 59 .000 Dress – Sweater .67 .80 .10 6.49 59 .000 Dress – Coat .43 1.00 .13 3.34 59 .001 Dress – Skirt .18 .72 .09 1.96 59 .055 Dress – Pants -.17 .67 .09 -1.36 59 .180 Dress – Top .53 .91 .12 4.54 59 .000

Table 3.3. Paired Sample t-Test Results for the Comparisons among Apparel Categories

in Importance of Touch and Fit

Pretest 2

The second pretest was conducted to choose two stimuli to present in a mock website and a mock catalog for the main experiment. A total of ten dresses were selected from commercial websites, which targeted young female consumers and which provided catalogs, and an online survey was developed including ten dresses. By using Adobe

Photoshop, the pictures of the ten dresses were manipulated in order to contain the same background, resolution, brightness, and contrast. A URL for the online survey was emailed to 54 college students who were enrolled in a class in the Textiles and Clothing program at the Ohio State University. A convenience sample of 32 female college students participated in the pretest 2 and randomly selected participants received a $10 gift certificate of a fashion brand as an incentive.

Ten items were evaluated in terms of attractiveness (1: highly unattractive and 7: highly attractive), fashionability (1: highly unfashionable and 7: highly fashionable), likableness (1: highly unlikable and 7: highly likable), and likelihood of purchase (1: highly unlikely to purchase and 7: highly likely to purchase) (See Appendix C). Since the four scale items were reliable for each dress (all Cronbach α’s >.7), scores from the four items were summed and used to select two apparel stimuli. Two apparel items with medium mean scores (the fifth and sixth ranks) were selected for the main study and presented across eight experimental conditions (See Table 3.4).

A t-test was used to assess whether the two dresses had similar levels of appealingness and revealed that there was no significant differences in appealingness between the two dresses, t (61) =.15, p = .88 (See Table 3.5). Since the selected apparel

items were presented on models, the attractiveness of the models was also examined in order to ensure no differences in attractiveness of the models who wore the two apparel items selected. Thus, the attractiveness of the ten models was also rated by Pretest 2 participants on the following characteristics: attractive/unattractive, beautiful/ugly, elegant/plain, and sexy /not sexy. Reliability of the four items was calculated for each

model and found to be reliable (all Cronbach α’s >.9). Therefore, the scores of four items

were summed. The results of a t-test showed that there were no differences in the

attractiveness of the two models who wore the two apparel stimuli, t (61) =.04, p = .97

(See Table 3.5).

Cronbach’s Mean SD Min. Max. alpha Apparel stimuli Item 1 0.86 20.28 4.52 10 28 Item 2 0.79 18.06 4.42 11 28 Item 3 0.93 14.94 6.29 5 28 Item 4 0.87 21.66 4.87 11 28 Item 5 0.92 13.66 7.41 4 25 Item 6 0.86 17.00 4.96 8 25 Item 7 0.93 15.09 7.36 4 28 Item 8* 0.96 16.50 7.74 4 28 Item 9 0.93 12.13 6.56 4 28 Item 10* 0.92 16.77 7.16 4 28 Model Model1 0.92 23.88 3.61 17 28 Model2 0.91 23.53 3.72 16 28 Model3 0.95 17.88 5.96 6 28 Model4 0.92 22.97 4.45 13 28 Model5 0.94 17.72 6.81 4 28 Model6 0.94 20.25 5.14 12 28 Model7 0.94 23.00 4.58 12 28 Model8* 0.94 18.34 5.89 6 28 Model9 0.94 15.34 6.07 4 28 Model10* 0.94 18.29 5.85 4 28 Note. * Items selected for the main study.

Table3.4. Ratings of Apparel Stimuli

Mean SD t p Apparel stimuli (IV) Appealingness (DV) Item 8 16.50 7.74 .146 .884 Item 10 16.77 7.16 Model (IV) Attractiveness (DV) Model 8 18.34 5.89 .036 .971 Model 10 18.29 8.85

Table 3.5. t-tests for the apparel stimuli and models

Pretest 3

Pretest 3 was conducted to assure that two sets of pictorial information (picture

swatch or no picture swatch) had different haptic imagery-provoking ability. Two

different online surveys were developed: the first online survey provided two pictures of

dresses with picture swatches and the other survey included only two pictures of dresses

without picture swatches (See Appendix D). An email was sent out to a convenient

sample of 102 college students who enrolled in two classes in the College of Human

Ecology, including one of two online surveys and explaining the purpose and confidentiality of the survey. As incentives all participants had extra credit for a class and randomly selected participants received a $10 gift certificate.

Participants were asked to complete a modified version of Ellen and Bone’s

(1991) Evoked Imagery Processing scale which included vividness and imagery

elaboration. A total of 67 students participated in the pretest, and 50 responses of female

students were used in the analysis of Pretest 3 after excluding 17 responses of males.

Twenty-five subjects evaluated product presentations with a picture swatch and twenty-

five participants viewed product presentations without the picture swatch. With 7-point

Likert scales (1: strongly disagree and 7: strongly agree), vividness was assessed by four

items (i.e., imagery which occurred was: clear, detailed, vague, vivid) and imagery

elaboration was measured through three items [When I looked at visual images, I imagined the feel of fabric textures of products (e.g., smooth/rough, flat/textured); I imagined what it would be like to touch the products; I fantasized that fabric properties of products (e.g., soft/hard, light/heavy)]. Since items assessing vividness and haptic imagery elaboration for two apparel stimuli were reliable (all Cronbach’s α > .7), each variable was summed. The t-test results indicated that product presentations with picture swatches had significantly higher vividness scores than product presentations without picture swatches. However, there was no significant difference in haptic imagery elaboration between the two product presentations although the means were in the right direction (See Table 3.6).

Pictorial Cronbach's N Mean SD t df p information alpha Dress1 Elaboration Swatch 25 15.64 4.01 .98 48 .333 .90 No swatch 25 14.64 3.17 Vividness Swatch 25 19.28 3.88 2.47 48 .017 .74 No swatch 25 16.56 3.89 Dress2 Elaboration Swatch 25 14.92 4.36 .80 48 .429 .93 No swatch 25 14.04 3.38

Vividness Swatch 25 20.00 3.63 2.40 48 .020 .78 No swatch 25 17.48 3.79

Table 3.6. Effects for visual information (picture swatch vs. no swatch) on vividness and haptic imagery elaboration

Pretest 4 (Content analysis)

Method

The purpose of Pretests 4 and 5 was to develop manipulations of verbal

information (high imagery and low imagery descriptions). In Pretest 4, a content analysis

was conducted in order to examine verbal and pictorial information of retail websites and

catalogs. Apparel websites were selected from Internet Retailer (2005) and

24hourmall.com. Internet Retailer (2005) published a list of the 400 largest retail websites identified by their 2004 sales and other statistical measures (e.g., monthly visits,

monthly unique visits, and sales conversation rate). 24hourmall.com is one of the most 59

popular shopping portal sites and provides the best and the most successful online retailers. A total of 47 apparel retailers providing both websites and catalogs were identified from Internet Retailer and 24hourmall.com and catalogs of the apparel retailers were requested. Since 36 apparel catalogs were received, 36 catalogs and the corresponding 36 websites were finally utilized in the content analysis with the goal of selecting a subset of dresses depicted in catalogs and on the corresponding websites at the same time.

A coding frame for the content analysis was developed based on past research on environmental elements of apparel retail websites (Ha et al., in press; Kim et al., 2006) and was refined according to the purpose of the present study. Information was categorized as verbal information or pictorial information. In order to compare information between online and catalog contexts, products (dresses) presented in both a website and the corresponding catalog of apparel retailers were selected. Two coders analyzed subsamples of five dresses selected from five apparel companies. First, coders found one dress displayed simultaneously in both a website and a catalog for each apparel brand. Since some apparel companies depicted only one dress both on their websites and in their catalogs at the same time, one dress for each apparel brand was used for the analysis. If more than one dress was identified from the website and the catalog of a company, among several dresses presented in both the website and the catalog, only the one dress which included the highest amount of information was selected. Next, two coders independently evaluated verbal and pictorial information for five dresses selected from five apparel companies. The information about dresses in online shopping and catalog shopping settings were separately coded. Inter-coder reliability was .92. After

five websites and five catalogs were analyzed, the initial coding structure was reviewed,

revised, and finalized. Subcategories of the coding frame were coded as ‘information is

present’ (1) or ‘information is absent’ (0).

Results

Verbal information

Price and size information was available on all 36 websites, but 35 catalogs

provided size information and 34 catalogs presented prices of products. Apparel retailers

which did not provide size (i.e., La Redoute, Ann Taylor) and price information (i.e., La

Redoute) in their catalogs focused on online shopping and used catalogs to present new

fashion items and themes of the season and brand concepts and images. Most websites

provided size charts (f=35), whereas size charts were available in 22 catalogs. In order to

provide concrete size information, some apparel retailers show measurements of body

parts for each size and explain how to measure each body part for the right fit by using

pictures. Since different apparel retailers use different sizing standards, detailed

measurement guides help shoppers to figure out the sizes and fit of products.

Country of origin was presented in 35 websites and 29 catalogs. Usually, the

websites and catalogs presented ‘imported’ instead of using the name of a specific

country (e.g., made in USA). Information about care instructions of a product (e.g., dry

clean, machine wash, hand wash) was provided by the majority of websites (83%) but

only by half of catalogs (53%).

Most websites (97%) and catalogs (89%) provided fabric and fiber content

information. The detailed fabric properties were available in some websites and catalogs.

Shape/structure (e.g., pique fabric, drape), texture (e.g., textured, smooth, silky), weight

(e.g., lightweight), temperature (e.g., cool), pressure (e.g., soft, flexible), and other fabric

details (e.g., embroidery, floral prints) were described. The detailed and concrete

information about fabric properties is useful because the information helps online and

catalog shoppers to perceive sensory properties of fabric and evaluate quality of fabrics

and products. In addition, some websites and catalogs try to stimulate shoppers’ past trial

experiences or imagination of product use by providing shoppers with descriptions related to contexts or events (e.g., wear between seasons or at the coast when cool

breezes blow in; our quick-drying Water Scooter went from beach cover-up to dinner

dress) and feelings (e.g., comfortable, fun).

Construction details and style information vary across websites and catalogs.

Since shoppers cannot try on products, apparel retailers try to provide accurate

information about construction and styles of products. For example, information about

length of products (e.g., falls at knee, 45” from natural waist), shape/silhouette (e.g., A-

line shape, full A-line skirt), and fit (semi-fitted, slim-fitting) help a shopper visually

imagine the configuration of products and the fit of products to his or her body. Since pictures presented in websites and catalogs are limited in describing design details of products, information about design details (e.g., dart/dart equivalents, closures, sleeves, necklines, hems) is useful for shoppers to assess product quality and reduce perceived risk.

Pictorial information

All 36 websites and 35 catalogs presented front views of products. Websites also provided additional views, such as larger views (f=19), zoom function (f=17), back views

(f=4), and side views (f= 4), whereas only two catalogs showed additional back views of a product. As compared to catalogs, websites allow consumers to simulate the use of products by using a mouse and clicking on a product image or a button that reads ‘back view’ or ‘alternative views’ (Ha et al., in press). The websites’ interactive features play a role in improving shoppers’ experiences in virtual spaces and mental images of the virtual world (Schlosser, 2003).

Pictures of fabric swatches and color swatches were available in some websites

(50%) and catalogs (22%). In particular, fabric swatches and close-up fabric swatches effectively present haptic properties of products (e.g., texture of velvet, smoothness of silk, texture of eyelet embroidery). When online and catalog shoppers look at fabric swatches, they may use visual clues to imagine the feeling of touching the fabrics and products. Concrete and vivid pictures of fabric swatches may enhance visual imagery and cognitive elaboration of tactile properties of a fabric.

Products were usually displayed on a model in both websites (89%) and catalogs

(89%). Product presentations on a model might be more useful in terms of providing information about the fit of product as compared to products presented on a hanger or flat.

Similar to verbal descriptions of contexts and events, background color and color surrounding a product may increase mental images of products. Websites (72%) and catalogs (72%) used contexts (e.g., summer beach, outdoor or building scenery) rather than solid colors (e.g., white, mild, black/dark colors) as background color and color

surrounding the product. Those contextual features may be effective in elaborating fantasies and enhance mental imagery.

Based on the results of the content analysis, the types of verbal information most frequently found on websites were used to create a list of verbal descriptions of dresses.

The verbal descriptions were evaluated in the next pretest to develop high haptic imagery and low haptic imagery descriptions.

Website Catalog Verbal information f % f % Fiber/fabric content 35 97.22 32 88.89 shape/structure 13 36.11 11 30.56 texture 3 8.33 3 8.33 Fabric/ Fabric weight 3 8.33 2 5.56 Fiber properties temperature 2 5.56 2 5.56 pressure 11 30.56 11 30.56 Other fabric details 15 41.67 15 41.67 Length 25 69.44 17 47.22 Shape/silhouette 15 41.67 12 33.33 Fit 10 27.78 6 16.67 Construction dart/dart equivalents 8 22.22 5 13.89 Details closures 16 44.44 14 38.89 (Style) Design sleeves 5 13.89 3 8.33 details necklines 17 47.22 14 38.89 hem 11 30.56 7 19.44 others 20 55.56 11 30.56 Color 25 69.44 22 61.11 Size chart 35 97.22 22 61.11 Size 36 100.00 34 94.44 Price 36 100.00 35 97.22 Care 30 83.33 19 52.78 Country of origin 35 97.22 29 80.56 Context cooperation (related to events or occasions) 9 25.00 9 25.00 Feelings 8 22.22 6 16.67

Continued

Table 3.7. Verbal and Pictorial Information Coded from Apparel Websites and Catalogs

Table 3.7. continued

Website Catalog Pictorial information f % f % front view 36 100.00 35 97.22 Types of back view 4 11.11 2 5.56 product side view 4 11.11 1 2.78 view larger view 19 52.78 0 0.00 zoom function 17 47.22 0 0.00 fabric swatch 10 27.78 7 19.44 Swatch color swatch 8 22.22 1 2.78 close-up fabric swatches 1 2.78 0 0.00 automatic color change 5 13.89 0 0.00 Product Color products with all colors on presentation presentation 1 2.78 2 5.56 same page hanging 0 0.00 1 2.78 mannequin 5 13.89 1 2.78 Product folding 0 0.00 0 0.00 display flat 9 25.00 4 11.11 method model 32 88.89 32 88.89 cyber model 1 2.78 0 0.00 Mix and suggestion for each item 6 16.67 1 2.78 match interactive mix & match 16 44.44 5 13.89 white 35 97.22 26 72.22 mild 3 8.33 4 11.11 Background black/dark 0 0.00 0 0.00 color vivid 1 2.78 1 2.78 context 0 0.00 5 13.89 Environment white 9 25.00 5 13.89 Color mild 4 11.11 5 13.89 surrounding black/dark 3 8.33 0 0.00 product vivid 1 2.78 1 2.78 context 26 72.22 26 72.22

Pretest 5

Based on the results of verbal information in Pretest 4, fifty phrases were selected

from descriptions of dresses presented on commercial apparel websites and catalogs. A convenient sample of college students (n=128) enrolled in a class in the Textiles and

Clothing program was used. A total of 103 female college students participated in the pretest. As incentives for participation, all participants were promised to receive extra course credit, and a $10 gift certificate of a fashion brand was provided to two randomly selected participants. They were given descriptions of dresses and fabric properties and were asked to evaluate the extent to which each description stimulates visual images or feelings of touch. Using a 5-point Likert scale, they rated each description’s haptic imagery-provoking ability [1: the description very slightly stimulates tactile (haptic) information and touch feelings (haptic imagery-provoking) and 5: very strongly stimulates tactile (haptic) information and touch feelings (non haptic imagery- provoking)] and visual imagery-provoking ability [1: the description very slightly stimulates visual images (visual imagery-provoking) and 5: very strongly stimulates visual images (visual imagery-provoking)].

Haptic imagery Visual imagery Descriptions of dresses Mean SD Mean SD Silky 4.52 0.80 4.22 0.92 Comfortable dress feels soft next to your skin 4.37 0.76 4.21 1.08 Luxuriously soft 4.33 1.00 3.54 1.28 Made of 100% combed cotton; cashmere 4.29 0.90 3.63 1.17 Smooth 3.85 1.11 3.40 1.12 Made of 55% Linen, 45% cotton; lycra 3.75 1.05 3.30 1.12 Fine/coarse 3.68 1.16 2.96 1.08 Stretch 3.66 1.18 3.45 1.13 Pique; crinkle fabric 3.62 1.16 3.80 1.03 Lightweight; featherweight 3.62 1.09 3.11 1.16 Wear between seasons or at the coast when cool 3.49 1.17 3.50 1.10 breezes blow in Beautifully textured cable 3.48 1.26 3.60 1.30 Eyelet texture 3.41 1.32 3.44 1.36 Softly gathered skirt 3.41 1.20 3.81 0.99 Flexible 3.41 1.19 3.21 1.15 Crisp 3.38 1.28 3.19 1.25 Glittering rhinestones 3.30 1.27 4.31 1.08 Embroidered 3.23 1.18 3.54 1.18 Floral beaded 3.17 1.27 3.75 1.25 Thin 3.09 1.65 3.25 1.31 Softly rounded hem 3.07 1.16 3.49 0.94 Fitted; feminine slim-fitting 3.06 1.32 4.33 0.84 Spaghetti straps 2.98 1.41 4.11 1.04 The most flattering style you will ever own 2.97 1.59 3.81 1.42 Fully lined 2.96 1.31 3.13 1.26 It never fails to get you noticed 2.92 1.51 3.88 1.29 Fresh and feminine summer style 2.89 1.41 3.42 1.12 Tight fit; slim fit 2.87 1.30 4.03 1.02 Adjustable straps 2.87 1.38 3.54 1.24 Cool 2.81 1.36 2.84 1.26 Flimsy 2.76 1.20 3.00 1.22 Scoop neckline; V neckline 2.73 1.39 4.27 0.96 Loose/compact 2.72 1.12 3.24 1.05 Contrast ribbon trims 2.71 1.27 3.69 1.13 Bracelet sleeves; short-sleeved; sleeveless 2.70 1.17 3.93 1.13 Not bulky 2.67 1.41 3.08 1.34

Continued

Table 3.8. Rating for descriptions of dresses

Table 3.8. continued

Twelve tightly placed snaps 2.66 1.29 3.53 1.25 Drapery; slouchy 2.62 1.19 3.13 1.35 Empire waist with bust darts 2.62 1.36 3.86 1.24 Flat 2.62 1.14 3.08 1.10 A-line shape 2.61 1.37 4.18 1.05 Smocked bodice 2.59 1.21 3.15 1.32 Machine wash, hand wash 2.56 1.35 2.62 1.40 Falls at knee 2.55 1.30 3.96 1.13 Lagoon blue, bright navy, sea rose 2.53 1.36 4.26 0.96 Choose from two styles in seven spring colors 2.45 1.53 3.72 1.40 Extra small, small, medium, large, extra large 2.44 1.44 3.57 1.34 Batik-print; printed with butterflies 2.44 1.36 3.69 1.50 Imported 2.05 1.51 2.11 1.41 $118; $69 2.00 1.24 2.50 1.42

Based on the participants’ responses about descriptions of dresses, high haptic imagery and low haptic imagery descriptions for two apparel stimuli which were selected in Pretest 2 were developed. The high imagery verbal information contained descriptions which were perceived as provoking high imagery (e.g., softness, lightness, silkiness, texture, and smoothness). The low imagery verbal information included information about design and construction details, or care instructions (e.g., adjustable straps, and bracelet sleeves). Both experimental conditions provided the same basic product information (e.g., color, price, size) (See Appendix E).

To ensure that the two high haptic imagery descriptions were different from the two low haptic imagery descriptions on their haptic imagery-provoking ability while not differing on persuasiveness and believability, two sets of verbal descriptions were rated in terms of haptic imagery representing 14 material properties, persuasiveness and

believability (See Table 3.9). Using 7-point Likert scales ranging from 1 (strongly disagree) to 7 (strongly agree), the 14 material properties were evaluated using the following sentence stem: ‘From the descriptions, I got a sense of the following garment and fabric aspects (e.g., smoothness, silkiness, texture).’ Participants were recruited from a class (n=56) of the Textiles and Clothing program and a total of 38 female college students participated in the pretest. For the first set of verbal descriptions (for the first dress), t-tests revealed significant effects with the high imagery information rated higher on smoothness, silkiness, texture, softness, flimsiness and drapability than low haptic imagery information. For the second set of verbal descriptions (for the second dress), t- tests showed significant differences in smoothness, silkiness, softness, lightness, thinness, and coolness between high imagery and low imagery verbal information. The high and low haptic imagery versions of verbal information did not differ on persuasiveness and believability (all p-values > .1).

Finally, two sets of verbal descriptions were refined so that two high haptic imagery descriptions contained information about 7 material properties (i.e., smoothness, silkiness, texture, softness, lightness, coolness, drapability) that were significant among the 14 properties. Flimsiness and thinness were excluded for the final manipulations since they may be associated with somewhat negative feelings and evaluations about products.

Additionally, reliability of the 7 items (material properties) was calculated for the two apparel stimuli. Since Cronbach’s alphas were greater than .9, composite scores were calculated by summing the item scores. Results from t-tests revealed that there were significant differences between high haptic imagery descriptions and low haptic imagery

descriptions for the single (summed) haptic indicator, t (36) = 3.23, p = .00; t (36) = 2.92, p = .01 (See Table 3.10).

Properties Verbal descriptions N Mean SD t df p Dress 1: Cotton tier dress Smoothness High haptic imagery 17 6.00 0.71 2.73 35 0.01 Low haptic imagery 20 5.05 1.28 Silkiness High haptic imagery 17 5.65 1.00 3.08 36 0.00 Low haptic imagery 21 4.52 1.21 Texture High haptic imagery 17 5.82 0.95 3.12 36 0.00 Low haptic imagery 21 4.81 1.03 Limpness High haptic imagery 17 5.29 1.31 1.71 36 0.10 Low haptic imagery 21 4.62 1.12 Softness High haptic imagery 17 6.06 0.97 3.17 36 0.00 Low haptic imagery 21 4.90 1.22 Flimsiness High haptic imagery 17 5.59 1.28 2.33 36 0.03 Low haptic imagery 21 4.67 1.15 Compactness High haptic imagery 17 5.53 1.23 1.21 36 0.23 Low haptic imagery 21 5.05 1.20 Flexibleness High haptic imagery 17 5.41 1.12 0.88 36 0.39 Low haptic imagery 21 5.10 1.09 Lightness High haptic imagery 17 5.94 0.97 1.75 36 0.09 Low haptic imagery 21 5.29 1.27

Continued

Table 3.9. Effects for verbal information (haptic imagery vs. non haptic imagery descriptions) on haptic imagery related to material properties

Table 3.9. continued

Bulkiness High haptic imagery 17 4.65 1.50 0.65 36 0.52 Low haptic imagery 21 4.33 1.46 Thinness High haptic imagery 17 5.59 1.23 1.57 36 0.12 Low haptic imagery 21 4.90 1.41 Drapability High haptic imagery 17 5.88 0.93 2.51 36 0.02 Low haptic imagery 21 4.86 1.46 Stretchiness High haptic imagery 17 5.12 0.93 1.36 36 0.18 Low haptic imagery 21 4.57 1.43 Coolness High haptic imagery 17 6.06 1.39 1.94 36 0.06 Low haptic imagery 21 5.14 1.49 Believability High haptic imagery 17 5.94 1.25 1.38 36 0.18 Low haptic imagery 21 5.33 1.43 Persuasiveness High haptic imagery 17 5.76 0.97 1.69 36 0.10 Low haptic imagery 21 5.00 1.64 Dress 2: Tiered floral dress Smoothness High haptic imagery 17 5.82 0.95 3.09 36 0.00 Low haptic imagery 21 4.62 1.36 Silkiness High haptic imagery 17 5.59 1.06 2.73 36 0.01 Low haptic imagery 21 4.52 1.29 Texture High haptic imagery 17 5.41 1.00 1.60 36 0.12 Low haptic imagery 21 4.76 1.41 Limpness High haptic imagery 17 5.41 0.51 2.05 36 0.05 Low haptic imagery 21 4.67 1.43 Softness High haptic imagery 17 5.88 0.60 2.90 36 0.01 Low haptic imagery 21 4.86 1.35 Flimsiness High haptic imagery 17 5.29 0.92 1.83 36 0.08 Low haptic imagery 21 4.57 1.40 Compactness High haptic imagery 17 5.47 1.18 1.40 36 0.17 Low haptic imagery 21 4.86 1.46 Flexibleness High haptic imagery 17 5.29 0.92 1.31 36 0.20 Low haptic imagery 21 4.71 1.62

Continued

Table 3.9. continued

Lightness High haptic imagery 17 5.82 0.64 2.17 36 0.04 Low haptic imagery 21 5.05 1.36 Bulkiness High haptic imagery 17 4.53 1.28 0.92 36 0.37 Low haptic imagery 21 4.10 1.58 Thinness High haptic imagery 17 5.35 0.86 2.72 36 0.01 Low haptic imagery 21 4.29 1.42 Drapability High haptic imagery 17 4.94 1.43 0.94 36 0.35 Low haptic imagery 21 4.52 1.29 Stretchiness High haptic imagery 17 4.94 1.20 0.28 36 0.78 Low haptic imagery 21 4.81 1.57 Coolness High haptic imagery 17 6.24 0.75 3.20 35 0.00 Low haptic imagery 20 4.95 1.50 Believability High haptic imagery 17 5.71 0.92 0.69 36 0.49 Low haptic imagery 21 5.43 1.43 Persuasiveness High haptic imagery 17 5.24 1.35 0.45 35 0.65 Low haptic imagery 20 5.00 1.75

Chronbach’s Verbal descriptions N Mean SD t df p alpha Dress1 High haptic imagery 17 41.41 5.51 3.23 36 0.00 .94 Low haptic imagery 21 34.33 7.55 Dress2 High haptic imagery 17 39.71 4.43 2.92 36 0.01 .94 Low haptic imagery 21 33.05 8.49

Table 3.10. Effects for verbal information (haptic imagery description vs. non haptic imagery description) on haptic imagery

CHAPTER 4

MAIN EXPERIMENT

This chapter presents the research method used to accomplish the purposes of this

study. This phase of the present study focuses on designing an experiment, developing a

mock website and a mock catalog, and explaining the sample and procedure.

Experimental Design

The proposed model and hypotheses were examined in an experimental study.

The design of this study was a 2 x 2 x 2 x 2 between-subjects factorial design.

Independent variables were pictorial information (picture swatch or no swatch), verbal information (high haptic imagery description or low haptic imagery description), shopping context (online shopping or catalog shopping), and NFT (high NFT or low

NFT). The forth independent variable, NFT, was measured; participants were divided into two groups based on a median split. Eight different experimental conditions were developed by using the three manipulated independent variables (i.e., pictorial information, verbal information and shopping context) (See Table 4.1). Subjects were randomly assigned one of eight experimental conditions.

The first and second experimental factors, pictorial and verbal information, were manipulated in two conditions: high haptic imagery or low haptic imagery conditions.

Based on Paivio’s dual-coding hypothesis, high imagery conditions are considered to increase probability that both imaginal and verbal processes play a role in memory and learning, such as concrete words and vivid pictures (Roeckelein, 2004). In the non-store shopping contexts accessing information about material properties of products (e.g., texture, weight, temperature) depends on pictures and verbal descriptions of products since products in that format can only be explored visually (without touch). Thus, in this study, high haptic imagery conditions are those with pictorial and verbal information which facilitate the imagination of apparel and fabric texture (e.g., silkiness, smoothness),

pressure (e.g., softness, flexibility), weight (e.g., lightness), and temperature (e.g.,

coolness).

Four different experimental conditions (High pictorial and high verbal

information, high pictorial and low verbal information, low pictorial and high verbal

information, and low pictorial and low verbal information) were presented in two

different non-store shopping contexts: online or catalog. In both the experimental

conditions of online and catalog shopping, participants engaged in a shopping simulation

to browse a website or a catalog and to purchase a product. Although consumers cannot

directly feel and touch products in either online or catalog shopping contexts, online

shopping involves a certain interactivity, defined as the extent to which users can

manipulate the form and content of a mediated environment in real time (Keng & Lin,

2006; Schlosser, 2003; Vijayasarathy & Jones, 2000). Therefore, the online shopping

condition in this study provides a function that allows shoppers to use mouse to click a

picture swatch to see a closer view and to navigate freely backward and forward through webpages.

Shopping contexts Internet shopping Catalog shopping Pictorial information Pictorial information Picture No swatch Picture No swatch swatch swatch High Internet Internet Catalog Catalog haptic shopping + shopping + shopping + shopping + Picture swatch No swatch Picture swatch No swatch imagery + High haptic + High haptic + High haptic + High haptic Verbal imagery imagery imagery imagery information Low Internet Internet Catalog Catalog haptic shopping + shopping + shopping + shopping + Picture swatch No swatch Picture swatch No swatch imagery + Low haptic + Low haptic + Low haptic + Low haptic imagery imagery imagery imagery

Table 4.1. Manipulated experimental conditions for the study

Mock Website and Catalog Development

By using pictorial information (picture swatch or no picture swatch) and verbal

information (high haptic imagery or low haptic imagery descriptions) developed in the

previous pretests, eight experimental conditions (four mock websites and four mock

catalogs) were developed. Both mock catalogs and mock websites simulated an apparel

retailer selling women’s apparel and targeting young women and contained a brand name

“Sky Shopping” and two apparel stimuli.

Participants were allowed to simulate a shopping process: browsing for the

information about two apparel stimuli, deciding a purchase of an item, and answering a

questionnaire. As compared to the mock catalogs, mock websites were developed to

include features which improve interactivity. The mock websites allowed participants to

freely navigate webpages by using hyperlinks. In the experimental condition with the picture swatch, participants were able to browse for a closer view of the picture swatch

by clicking a picture swatch icon. On the other hand, mock catalogs were presented in

two catalog pages. Since the mock catalogs simulated print catalogs, participants were

able to browse two catalog pages by simply scrolling down without closer view functions.

Thus picture swatches were simply presented on the catalog pages to simulate actual

catalog pages (See Appendix F).

Sample and Procedure

Firstly, this research was reviewed and exempted by an Ohio State University

Institutional Review Board (IRB) and was assigned a protocol number 2006E0302

(Appendix H). The sample for the main experiment was a random sample of Ohio State

University female students. Previous research reported characteristics of nonstore shoppers. The general characteristics of nonstore shoppers are younger, well-educated, technology oriented, and convenience oriented (Lu & Rucker, 2006). It has been reported that young, unmarried women and students tend to have favorable attitudes toward catalog shopping (Sherman, 1996) and women are the main patrons of apparel and beauty websites (Internet Retailer, 2004). Online and catalog apparel retailers are targeting young female consumers. For that reason, female college students were used as a sample of the study. The Office of the University Registrar randomly extracted 6,400 female student names from all female college students enrolled at the Ohio State University for the quarter and provided a list of email addresses for the study.

Participants were randomly assigned to one of eight experimental conditions: (1) high haptic imagery description /picture swatch/online shopping context, (2) high haptic description/no swatch/online shopping context, (3) low haptic imagery description/picture swatch/online shopping context, , (4) low haptic description/no swatch/online shopping context, (5) high haptic imagery description /picture swatch/catalog shopping context, (6) high haptic description/no swatch/catalog shopping context, (7) low haptic imagery description/picture swatch/catalog shopping context, and

(8) low haptic description/no swatch/catalog shopping context. Potential participants

received an invitation letter via email which explained the purpose of the study,

confidentiality of information obtained from the research, an incentive for participants of

the study, and a URL address which was linked to one of the eight experimental

conditions. Thus, each experimental condition was assigned to 800 potential participants.

As an incentive for participation, diverse levels of prizes were promised with a random

drawing: one winner (a $100 gift certificate), two winners (a $50 gift certificate), and

fifteen winners (a $20 gift certificate). Two reminders were emailed. Seven days after the

initial invitation message was emailed, the first reminder email was sent to those who had

not participated. Then, seven days later, the second reminder was also emailed.

Participants who clicked on the survey URL were first asked to type their name to indicate their agreement to participate in the study and then to move to the experimental website or catalog. If respondents did not want to participate in the experiment and to complete a questionnaire, they were allowed to exit the process.

Before browsing a mock website or a mock catalog, participants were led to a welcome page which included instructions and a scenario for the experiment in order to set up a high involvement condition. Experimental conditions designed for the online shopping context provided the following scenario: “One day you are browsing the

Internet. You have been given a $150 gift certificate to make a purchase from a website, www.Skyshopping.com. Recently, you have recognized the need for a dress. You are planning to purchase a dress with the gift certificate from the website.” On the other hand, the scenario for catalog shopping context was as follows: “One day you have received a catalog. You have been given a $150 gift certificate to make a purchase from

the catalog, Skyshopping. Recently, you have recognized the need for a dress. You are

planning to purchase a dress with the gift certificate from the catalog.”

Next, participants were instructed to browse the catalog or the website and then to decide on an apparel item, and to complete the measures of perceived product quality, perceived risk, attitude toward a product, behavioral intentions, NFT, and demographics.

The online survey responses were collected during four weeks. Winners for the random

drawing were noticed by email. The email contained a congratulation letter and an

acceptance form of a prize for the survey participation. Winners who replied to the email

and filled in the acceptance form were able to receive the prize via mail. If winners did

not respond to the email within a week, other winners were selected.

Measures

Dependent measures used in the main experiment were (1) perceived product

quality, (2) perceived risk, (3) attitude toward a product, and (4) behavioral intentions.

After subjects browsed stimulus websites or catalogs, they were asked to complete a

questionnaire including the four dependent measures and NFT. In order to gather general

information about participants, demographic items and items about use of the Internet

were included (See Table 4.2 and Appendix G).

Perceived Product Quality

Perceptions of product quality were measured by assessing 14 attributes of a

garment with 7-point Likert-type scales ranging from 1 (low quality) to 7 (high quality).

Davis (1985) used 8 items to measure perceptions of clothing quality and reported the reliability of the scale (Cronbach’s α = .87). Eckman et al. (1990) identified 14 criteria for women’s general garment evaluations. In this study, 14 attributes for the evaluation of a garment in the current study were used including color/pattern, style, fabric, uniqueness, appearance, versatility, matching, appropriateness, utility, fit, comfort, care, workmanship, and price.

Perceived Risk

The perceived risk scale of Kim and Lennon (2000) and Park (2002) were used.

Park revised Kim and Lennon’s scale developed for television shopping in order to

measure perceived risk in online shopping context. Kim and Lennon (2000) identified

three factors with 13 items: uncertainty about products (Cronbach’s α = .91), negative

attitude toward television shopping (Cronbach’s α = .87), uncertainty about

consequences (Cronbach’s α = .74). Park and Stoel (2005) reported overall reliability

(Cronbach’s α = .91). The 24 item Likert scale was used with endpoints “strongly agree

(7)” and “strongly disagree (1).” For catalog shopping, the scale was revised. For

example, ‘Buying the dress from the website that I saw today is risky because size may

not fit me’ was changed to ‘Buying the dress from the catalog that I saw today is risky

because size may not fit me.’

Attitude toward a Product

Attitudes toward a product were measured by six items with 7-point semantic

differential scale: the dress is ‘bad-good’, ‘unappealing-appealing’, ‘unpleasant- pleasant’,

‘unattractive-attractive’, ‘boring-interesting’, and ‘dislikable-likable’. The attitude scale was developed by Bruner (1998) and reliability of the scale was reported by Li,

Daugherty, and Biocca (2002) (Cronbach’s α = .91).

Behavioral Intentions

Behavioral intention is an effective measure to anticipate a response behavior.

Past researchers developed behavioral intention scales (Bloemer, Ruyter, & Wetzels,

1999; Srinivasan, Anderson, & Ponnavolu, 2002; Zeithaml et al., 1996) and reported

reliability of the scales (all Cronbach’s > .7). Three items of behavioral intention in this study were measured including (1) purchasing an apparel item, (2) paying extra money for the item, and (3) recommending the website (or catalog) to my friends and family.

The established behavioral intention items were measured using 7-point semantic differential scales: ‘unlikely-likely’, ‘improbable-probable’, and ‘impossible-possible’

(Kwon, 2005). Kwon also provided reliability of the scale for online shopping context

(Cronbach’s α = .90) and offline shopping context (Cronbach’s α = .85).

Need for Touch (NFT)

The Need for Touch (NFT) scale of Peck and Childer (2003b) was used to measure individual differences in motivation to acquire and use haptic information. All

12 items were measured using 7-point Likert scales ranging from 1 (strongly disagree) to

7 (strongly agree). The reliability of the NFT Scale was reported (Cronbach’s α = .95)

(Peck & Childers, 2003b). Examples are ‘I feel more comfortable purchasing a product after physically examining it’ and ‘If I cannot touch a product in the store, I am reluctant to purchase the product.’

Variables Items Cronbach’s alpha Perceived product 1. Color/pattern .87 quality 2. Style : Davis, 1985; 3. Fabric Eckman, Damhorst, 4. Uniqueness & Kadolph, 1990 5. Appearance (Attractiveness) 6. Versatility (Various end use) 7. Matching 8. Appropriateness 9. Utility 10. Fit 11. Comfort 12. Care 13. Workmanship (construction) 14. Price 15. Overall quality

Perceived risk: Kim 1. The color may not be what I thought it would be. Overall (.91); & Lennon, 2000; 2. Size may not fit me. ;physical risk (.83) Park, 2002; Park & 3. There may be something wrong with the apparel purchased (e.g., ;functional risk Stoel, 2005 broken button, damaged fabric). (.85) 4. I may want to exchange it for another item. ;social/psychologi 5. I may not like it. cal risk (.80) 6. It may not look good on me. ;economic/financi 7. My friends may think I look funny when I wear it. al/privacy risk 8. I may not be able to match it with my current clothing. (.79) 9. I may not feel comfortable wearing it in public. 10. I may have to pay for an alteration (i.e., lengthen or shorten the hem). 11. It may be harmful to my health (chemical agent-allergic reason). 12. I may feel that I just threw away a lot of money. 13. I may feel that I just wasted time shopping via the Internet. 14. I may not feel comfortable giving my credit card number when I order. 15. The construction quality may be poor (e.g., poorly done stitches). 16. It may not be durable when cleaned (e.g., color changes, shape changes). 17. I may not wear the item. 18. I may find the very same item at the store with a lower price. 19. I may have a hard time trying to return or exchange the item. 20. If I return the item, I may not be able to get a full refund. 21. I may lose money if I purchase this apparel item (e.g., because it costs more than it should to keep it in good shape, because I will not be able to wear after one season). 22. There may be something wrong with this apparel, or it may not function properly (e.g., a raincoat will not be waterproof). 23. It may affect the way others think of me. 24. It may be a risky purchase.

Continued

Table 4.2. Items for dependent variables

Table 4.2. continued

Attitude toward a The dress is .91 product: Bruner, 1. good - bad 1998; Li, 2. unappealing - appealing Daugherty, & 3. unpleasant - pleasant Biocca, 2002 4. unattractive - attractive 5. boring - interesting 6. dislikable - likable

Behavioral 1. I would purchase the dress which I evaluated. Online: .90, intentions: Kwon, (1) unlikely – likely Offline: .85 2005 (2) improbably – probably (3) impossible – possible 2. I would be willing to pay extra in order to buy the dress. (1) unlikely – likely (2) improbably – probably (3) impossible – possible 3. I would recommend the website to my friends and family. (1) unlikely – likely (2) improbably – probably (3) impossible – possible

NFT: Peck & 1. Touching products can be fun. .95 Childers, 2003b 2. I place more trust in products that can be touched before purchase. 3. I like to touch products even if I have no intention of buying them. 4. I feel more comfortable purchasing a product after physically examining it. 5. When browsing in stores, I like to touch lots of products. 6. When walking through stores, I cannot help touching all kinds of products. 7. I feel more confident making a purchase after touching all kinds of products. 8. If I cannot touch a product in the store, I am reluctant to purchase the product. 9. The only way to make sure a product is worth buying is to actually touch it. 10. When I browsing in stores, it is important for me to handle all kinds of products. 11. I kind myself touching all kinds of products in stores. 12. There are many products that I would only buy if I could handle them before purchase.

CHAPTER 5

ANALYSIS AND RESULTS

This chapter presents research findings of the study. Sample characteristics and preliminary analyses are discussed first. Preliminary analyses address validity and reliability of measures, manipulation checks, and model specification. Exploratory and confirmatory factor analyses were performed to assess properties of measures.

Multivariate analyses of variance and structural equation modeling were employed to test proposed hypotheses. After discussing the results of proposed hypotheses, the chapter proposed an alternative model and discussed findings of the alternative model. SPSS,

CEFA, and RAMONA were used to analyze data.

Sample Characteristics

Of 6,400 female college students who were originally invited to participate in the experiment, 1,385 completed the research. The response rate for this study was 21.64%

(1385/6400). Since 30 respondents participated in one of the pretests and 56 others were missing the majority of responses, 86 respondents were excluded from the data analyses and finally a total of 1299 responses were used for the study. The participants for the

eight experimental conditions were 168 (picture swatch/high haptic imagery description/online shopping), 165 (picture swatch/low haptic imagery description/online shopping), 169 (no swatch/high haptic imagery/online shopping), 162 (no swatch/low haptic imagery/online shopping), 151 (picture swatch/high haptic imagery description/catalog shopping), 151 (picture swatch/low haptic imagery description/catalog shopping), 173 (no swatch/high haptic imagery/catalog shopping) and

160 (no swatch/low haptic imagery/catalog shopping) (See Table 5.1).

Experimental Pictorial Verbal Shopping Frequencies Percent Condition information information context Condition 1 picture swatch high haptic imagery online 168 12.9 Condition 2 picture swatch low haptic imagery online 165 12.7 Condition 3 no swatch high haptic imagery online 169 13.0 Condition 4 no swatch low haptic imagery online 162 12.5 Condition 5 picture swatch high haptic imagery catalog 151 11.6 Condition 6 picture swatch low haptic imagery catalog 151 11.6 Condition 7 no swatch high haptic imagery catalog 173 13.3 Condition 8 no swatch low haptic imagery catalog 160 12.3

Table 5.1 Participants for the Experimental Conditions

Frequency statistics provided sample characteristics of the study (See Table 5.2).

The average age of participants was 21 years. Ages ranged between 15 and 56 and the majority of participants (90%) were between 18 and 23 years old. About 80% of the participants were Caucasian (80.1 %), followed by African American (6.4%), Asian

(6.3%), and Hispanic (2.7%). The academic rank of the sample consisted of senior

(38.2%), sophomore (21%), junior (21%) and freshman (15.6%). The majors of participants were diversely distributed. Among various majors, Social and Behavioral

Sciences (14.7%), Human Ecology (13.9 %), Business (11.5%), Medicine/ Nursing/

Optometry/ Pharmacy (10.4 %) represented high portions of majors of participants.

Participants’ online and catalog shopping experiences were also examined by using 7-point Likert scales ranging from 1 (very infrequently) to 7 (very frequently).

Only those exposed to websites were asked about online shopping experience (N = 635), while only those exposed to catalogs were asked about catalog shopping experience (N =

664). Of 635 participants exposed to mock websites, about 78% of them very frequently use the Internet. However, they shop and purchase online less frequently compared to the general use of the Internet. Over half of them responded that they shop online (69%) and purchase online (58%) very frequently, frequently or sometimes. About 56% of participants shop for apparel online sometimes, frequently or very frequently and about

42% of them had experiences purchasing apparel online sometimes, frequently or very frequently. On the other hand, among 664 participants exposed to the mock catalogs, approximately half of them use catalogs very frequently, frequently or sometimes (53%).

Over 40% of participants shop catalog (44%) and shop for apparel from catalog (43%) very frequently, frequently, or sometimes. About one third of them purchase products

from catalogs (32%) very frequently, frequently or sometimes. Also about one third of them purchase apparel from catalogs (31%) very frequently, frequently or sometimes. In addition, many of them used catalogs very infrequently to shop. Participants responded that they very infrequently use catalogs (21%), very infrequently shop from catalogs

(25%), very infrequently purchase from catalogs (33%), and very infrequently purchase apparel from catalogs (36%) (See Table 5.3).

Variable Categories Frequencies Percent Age Under 18 3 .3 18-19 335 25.8 20-21 548 42.2 22-23 279 21.5 24-25 43 3.3 Over 25 89 6.9 Academic standing Freshman 202 15.6 Sophomore 273 21.0 Junior 273 21.0 Senior 496 38.2 Others 44 3.4 Ethnic background African American 83 6.4 Caucasian American 1040 80.1 Hispanic 35 2.7 Native American 4 0.3 Asian/Asian American 82 6.3 Others 40 3.1

Continued

Table 5.2. Sample Demographic Characteristics

Table 5.2. continued

Major Architecture 10 0.8 Art/Music 71 5.5 Biological Sciences 114 8.8 Business 149 11.5 Education 73 5.6 Engineering 49 3.8 Food, Agriculture & Environmental Sciences 47 3.6 Human Ecology 180 13.9 Humanities 124 9.5 Journalism & Communication 73 5.6 Law 3 0.2 Mathematical & Physical Sciences 16 1.2 Medicine/ Nursing/ Optometry/ Pharmacy 135 10.4 Social & Behavioral Sciences 191 14.7 Social Work 23 1.8 Veterinary Medicine 9 0.7 Others 11 0.8 Undecided 11 0.8

Online Online General Online Online apparel apparel Internet use shopping purchase shopping purchase N = 635 f % f % f % f % f %

(1) Very infrequently 27 4.3 50 7.9 80 12.6 103 16.7 157 24.7 (2) 6 .9 69 10.9 89 14.0 86 13.5 123 19.4 Infrequently (3) 4 .6 68 10.7 90 14.2 76 12.0 76 12.0 (4) Sometimes 24 3.8 136 21.4 110 17.3 86 13.5 77 12.1 (5) 18 2.8 122 19.2 120 18.9 100 15.7 86 13.5 Frequently (6) 49 7.7 85 13.4 67 10.6 83 13.1 47 7.4 (7) Very frequently 492 77.5 97 15.3 71 11.2 89 14.0 54 8.5 (0) Not applicable 15 2.4 8 1.3 8 1.3 9 1.4 11 1.7

Catalog Catalog General Catalog Catalog apparel apparel catalog use shopping purchase shopping purchase N = 664 f % f % f % f % f %

(1) Very infrequently 136 20.5 167 25.2 222 33.4 196 29.5 241 36.3 (2) 97 14.6 111 16.7 121 18.2 94 14.2 127 19.1 Infrequently (3) 67 10.1 83 12.5 94 14.2 73 11.0 69 10.4 (4) Sometimes 129 19.4 112 16.9 97 14.6 108 16.3 89 13.4 (5) 69 10.4 70 10.5 51 7.7 81 12.2 57 8.6 Frequently (6) 90 13.6 64 9.6 37 5.6 57 8.6 34 5.1 (7) Very frequently 62 9.3 43 6.5 29 4.4 39 5.9 28 4.2 (0) Not applicable 13 2.0 12 1.8 12 1.8 16 2.4 16 2.4

Table 5.3. Participants’ Online and Catalog Shopping Experience

Preliminary Analyses

Descriptive and Factor Analyses of Measurements

The study uses one moderating variable (need for touch) and four dependent

variables (perceived product quality, perceived risk, attitude toward a product, and

behavioral intention). In this section the descriptive statistics for the five variables (See

Table 5.6) and exploratory factor analyses of two variables (perceived product quality

and perceived risk) are presented.

Perceived Product Quality

Fourteen general garment evaluative criteria of Eckman et al. (1990) were used to

assess perceived product quality. Eckman et al.’s criteria were related to one of four

categories (aesthetic criteria, usefulness criteria, performance/quality, and extrinsic

criteria). Since the criteria were identified based on consumer interviews and frequency

distributions of the interviewers’ responses, an exploratory factor analysis (EFA) was

conducted in this study to identify dimensions of perceived product quality. Although a

one-dimensional scale with multiple items has been used to assess perceived apparel

quality in previous studies (Davis, 1985; Rao & Monroe, 1988), several researchers have emphasized the need for identifying multiple dimensions of perceived product quality

(Forsythe, Presley, & Caton, 1996: Stone-Romero et al., 1997).

Using the Comprehensive Exploratory Factor Analysis (CEFA) 2.0, Maximum

Likelihood (ML) model estimation and oblique rotation with CF-Quartimax criterion were employed to perform the EFA. Eckman et al. (1990) conceptually represented four

categories for evaluative criteria of apparel quality, but in this analysis two factor models were considered based on Kaiser’s Criterion (1960) and the Scree test (Cattell, 1966) because two eigenvalues were greater than 1.0. Therefore, 2, 3, and 4 factor solutions were examined. After excluding items with factor loadings which were lower than .40, factor models were re-examined. In each factor model, measures of model fit, such as

RMSEA (Root Mean Square Error of Approximation), were taken into consideration.

RMSEA estimates which are smaller than .05 indicate a close fit of the model, values of

RMSEA between .05 and .08 indicate a reasonable fit of the model, RMSEA estimates between .08 and .10 represent a mediocre fit of the model, and values of RMSEA which are greater than .10 indicate an unacceptable fit (Browne & Cudeck, 1992).

The EFA finally yielded a 3-factor model with 10 items (See Table 5.4). The

RMSEA estimate indicated a reasonable fit (.058) for the model. The first factor,

‘usefulness’, addressed consumers’ evaluations about the quality of mix-and-match potential, diverse uses, and appropriateness for social settings. Four items (i.e., versatility, matching, appropriateness, and utility) were identified for the first factor. The second factor was labeled as ‘performance quality’ related to concerns about fit, comfort and care of a garment. The last factor, labeled as ‘aesthetic quality’, was associated with the physical appearance and aesthetics of a product and included three items (i.e., color/pattern, style, and appearance). Scores on the items within each factor were averaged and used as indicators of the latent variable of perceived product quality. The mean values of ‘usefulness’ for the overall sample was 5.03, the mean value of the second factor ‘performance quality’ was 5.23, and the mean value of the third factor

‘aesthetic quality’ was 5.50. The mean value of each factor for online and catalog

shopping contexts were separately calculated and are also presented in Table 5.6. The

Cronbach’s α for overall perceived quality was .90, Cronbach’s α for usefulness was .83

(α for online = .84, α for catalog = .81), Cronbach’s α for performance quality was .78 (α for online = .80, α for catalog = .77), and Cronbach’s α for aesthetic quality was = .85 (α for online = .86, α for catalog = .85) (see Table 5.6).

Factor loadings Items 1 2 3

Factor 1: Usefulness Q9 Utility .687 Q6 Versatility (various end uses) .672 Q7 Matching .601 Q8 Appropriateness .540

Factor 2: Performance quality Q11 Comfort .883 Q10 Fit .542 Q12 Care .538 Factor 3: Aesthetic quality Q2 Style .828 Q5 Appearance (attractiveness) .724 Q1 Color/pattern .685 Eigenvalue 5.21 1.11 .85 Sample discrepancy function value χ2(df) 100.040 (18) p for the test of exact fit .000 p for the test of close fit .119 RMSEA (C.I.) .058 (.047; .069)

Table 5.4. EFA Results for the 10item, 3-Factor Model for Perceived Product Quality

Perceived Risk

Twenty-four items of the perceived risk scale which were developed by Kim and

Lennon (2000) and revised by Park (2002) were used. Kim and Lennon identified three factors (i.e., uncertainty about the products, negative attitude toward television shopping, and uncertainty about consequences) of perceived risk in the television shopping context.

Park (2002) and Park et al. (2005) used summed scores of the scale with one dimension in their studies of online shopping. In order to determine possible dimensions of perceived risk, an EFA was conducted by employing ML model estimation and oblique rotation with CF-Quartimax criterion.

A four factor model with 12 items was identified by the EFA. The RMSEA estimate was .038, indicating a close fit of the model to the data for the model. The first factor was labeled as ‘social risk’ and addressed uncertainty about what others might be thinking about the consumer. The first factor included four items (e.g., My friends may think I look funny when I wear it). The second factor ‘financial risk’ referred to loss of money due to difficulties returning the item and included three items (e.g., I may find hard time try to return or exchange the item). The third factor was labeled as

‘psychological risk’ and items reflected the probability that the apparel is inconsistent with the consumer’s self-image or is aesthetically unsatisfactory. An example item from the third factor is ‘It may not look good on me.’ The fourth factor was ‘product performance risk’, which referred to the probability of the apparel failing to meet expected functions. Two items were loaded on the fourth factor and related to the poor quality of construction. Item scores within for a factor were averaged and higher scores reflected higher levels of perceived risk. The factors were used as manifest variables for

the latent perceived risk variable (See Table 5.5). The Cronbach’s α was .83 for overall

perceived risk, .80 for social risk (online = .81, catalog = .79), .78 for financial risk

(online = .79, catalog =.77), .78 for psychological risk (online = .79, catalog = .78) and .84 for product performance risk (online = .83, catalog = .86) (See Table 5.6). The

mean value of the first factor ‘social risk’ was 3.53, the mean value of the second factor

‘financial risk’ was 5.15, the mean value of the third factor ‘psychological risk’ was 5.69,

and the mean of the fourth factor ‘product performance risk’ was 4.65 (See Table 5.6).

Factor loadings Items 1 2 3 4 Factor 1: Social risk R7 My friends may think I look funny when I wear it. .751

R8 I may not be able to match it with my current .729 clothing. R9 I may not feel comfortable wearing it in public. .718

R23 It may affect the way others think of me. .633

Factor 2: Financial risk R19 I may find a hard time trying to return or .977 exchange the item. R20 If I return the item, I may not be able to get a full .720 refund. R18 I may find the very same item at the store with a .431 lower price.

Factor 3: Psychological risk R6 It may not look good on me. .904 R5 I may not like it. .624 R24 It may be a risky purchase. .597

Factor 4: Product performance risk R16 It may not be durable when cleaned (e.g., color .930 changes, shape changes). R15 The construction quality may be poor (e.g., poorly .760 done stitches). Eigenvalue 4.30 1.90 1.40 .98 Sample discrepancy function value .054 χ2(df) 70.020(24) p for the test of perfect fit p for the test of close fit .964 RMSEA (C.I.) .038 (.028;.049)

Table 5.5. EFA Results for the 12item, 4-Factor Model for Perceived Risk

Attitude toward a Product

The attitude scale (Bruner, 1998) was found to be reliable in this study

(Cronbach’s α for the overall sample = .92, α for online = .92 and α for catalog = .93).

Descriptive statistics of six items for attitude toward a product were displayed in Table

5.6.

Behavioral Intention

Three items were used to measure behavioral intention (See Table 5.6). Each item

(e.g., I would like to purchase the dress which I evaluated) was measured with three

semantic differential scale items (i.e., ‘unlikely-likely’, ‘improbable-probable’, and

‘impossible-possible’) and scores of three semantic differential items for each behavioral

intention question were averaged. Cronbach’s α for behavioral intention was calculated

(Cronbach’s α for the overall sample = .77, α for online = .75 and α for catalog = .78).

NFT

Need for touch (NFT) was measured by using 12 items developed by Peck and

Childer (2003b). Cronbach’s α of the scale was .92, indicating reliability (internal

consistency) of the scale. In order to divide the subjects into two groups regarding NFT,

12 items were averaged and a median-split was used to label people above the median

(Mdn = 5.36) as high NFT group, and those below the median as low NFT group. Among

the participants, 649 respondents (P = 50%) were classified as low in NFT and 650

respondents (P = 50%) were classified as high in NFT. There was a significant difference

in NFT between high NFT subjects (M = 6.19, SD = .47) and low NFT subjects (M =

4.35, SD = .83), t = 49.58, df=1297, p = .000.

Total sample Online Catalog

(N = 1299) (N = 635) (N = 664) Cronbach’s Cronbach’s Cronbach’s Min. Max. M SD M SD M SD α α α Perceived product quality PQ1 Usefulness 1 7 5.03 1.05 .83 5.02 1.11 .84 5.04 .99 .81 PQ2 Performance quality 1 7 5.23 1.07 .78 5.25 1.09 .80 5.22 1.05 .77 PQ3 Aesthetic quality 1 7 5.50 1.14 .85 5.50 1.16 .86 5.49 1.12 .85 Perceived risk PR1 Social 1 7 3.53 1.39 .80 3.51 1.41 .81 3.54 1.38 .79 PR2 Financial 1 7 5.15 1.30 .78 5.15 1.30 .79 5.16 1.29 .77 PR3 Psychological 1 7 5.69 1.14 .78 5.70 113 .79 5.68 1.14 .78 PR4 Product performance 1 7 4.65 1.42 .84 4.61 1.39 .83 4.68 1.44 .86 Attitude toward a product A1 Good/bad 1 7 5.70 1.24 .92 5.70 1.21 .92 5.69 1.27 .93 A2 Appealing/unappealing 1 7 5.47 1.52 5.51 1.47 5.44 1.56 A3 Pleasant/unpleasant 1 7 5.59 1.29 5.61 1.25 5.57 1.32 A4 Attractive/unattractive 1 7 5.58 1.45 5.59 1.43 5.56 1.46 A5 Interesting/boring 1 7 5.36 1.32 5.36 1.34 5.35 1.30 A6 Likable/dislikable 1 7 5.64 1.37 5.63 1.35 5.65 1.39 Behavioral intentions I1 I would like to 1 7 3.80 1.94 .77 4.01 1.78 .75 4.14 1.83 .78 purchase the dress which I evaluated. I2 I would like to pay 1 7 2.04 1.40 2.22 1.35 2.28 1.42 extra in order to buy the dress. I3 I would recommend 1 7 4.38 1.67 4.33 1.65 4.42 1.69 the website (catalog) to my friends and family. NFT N Need for touch 1 7 5.27 1.14 .92 5.29 1.14 .92 5.25 1.14 .92

Table 5.6. Descriptive Statistics and Reliability for Four Variables

100

Initial Model Considerations

According to Anderson and Gerbing’s (1988) two-step modeling approach, confirmatory factor analysis (CFA) was used to assess convergent and discriminant validity of the measurements. Using RAMONA (RAM Or Near Approximation),

Maximum Likelihood function was employed to estimate parameters for the CFA. In order to achieve convergent and discriminant validity, first the measurement model was evaluated and then respecified based on theoretical and statistical considerations.

Anderson and Gerbing (1988) provided four basic ways to respecify indicators: (1) relating a problematic indicator to a different factor, (2) erasing the indicator from the model, (3) associating the indicator with multiple factors, or (4) utilizing correlated measurement errors. According to the second way, the measurement model was respecified by deleting problematic indicators. Respecification was taken into consideration if an indicator contained a low factor loading (e.g., lower than .60) and a low squared multiple correlation (SMC) value (e.g., lower than .40), as compared to other indicators from the same factor (Bagozzi & Yi, 1988).

PR1 (perceived social risk) was deleted because the indicator had a low factor loading (.42) and a low SMC (.18) as compared to other indicators (PR2, PR3, and PR4) of perceived risk. Because A5 (the fifth indicator for attitude toward a product) had a low

SMC (.34) and A1 and A6 were related to other indicators which belonged to other latent variables (e.g., behavioral intentions and perceived product quality), A1, A5, and A6 were excluded from the final measurement model. Therefore, four indicators (i.e., PR1,

A1, A5, and A6) were deleted from the measurement model and 12 indicators were used for the final structural model. In addition, the errors of PQ1 (usefulness) and PQ2

101

(performance quality) were allowed to be correlated in the model because the

relationships between error variances of two indicators may be explained by other factors

(e.g, individual characteristics) but may be less associated with the other indicator

(aesthetic quality) from the same latent construct. For example, utilitarian shopping value may influence the evaluation of apparel in terms of both usefulness and performance.

Utilitarian shoppers may more consider usefulness, durability, and care of a product rather than aesthetic characteristics of the product. Therefore, based on the respecification process, the measurement model for the study was finalized and a final CFA model and factor loadings are presented in Figure 5.1 and Table 5.8.

The overall fit of the measurement model was assessed. Although the chi-square statistic was significant (χ2 = 456.89, df = 119, p = .000), indicating that the proposed

model failed to fit the data, other fit indices were considered because chi-square statistic

tends to be sensitive to large sample size (N=1299). The point estimate of RMSEA

was .047 with a confidence interval between .042 and .051, indicating a close fit of the

model to the data. The value of Tucker-Lewis Index (TLI) was .96, Normed Fit Index

(NFI) was .96, and Incremental Fit Index (IFI) was .97. The fit indices (TLI, NFI, IFI)

indicated the good overall model fit since they are greater than .95 (Bollen, 1989).

102

Θδ1,2

θδ1 θδ2 θδ3 θδ7 θδ8 θδ9

δ1 δ2 δ3 δ7 δ8 δ9

1 1 1 1 11

PQ1 PQ2 PQ3 A2 A3 A4

λ11 λ21 λ31 λ73 λ83 λ93

φ31

Perceived Attitude Product toward

Quality (ξ1) a Product (ξ3)

φ41 φ23 1 1 φ21 φ34 1 1

Perceived Behavioral Risk Intentions

(ξ2) (ξ4)

φ24

λ42 λ52 λ62 λ10,4 λ11,4 λ12,4

PR2 PR3 PR4 I1 I2 I3

1 1 1 1 1 1

δ4 δ5 δ6 δ10 δ11 δ12

Θδ4 Θδ5 Θδ6 Θδ10 Θδ11 Θδ12

Figure 5.1. A CFA Model for Final Measurement Items

103

Unidimensionality, convergent validity, and discriminant validity were assessed.

Unidimensionality refers to the existence of one latent construct underlying a set of measures (Anderson, Gerbing, & Hunter, 1987) and convergent validity refers to the degree that indicators of the same construct are highly correlated and show a uniform pattern of intercorrelations (Bagozzi, 1981). Unidimensionality was assessed by an exploratory analysis for each construct. Each exploratory factor analysis produced one factor solution and Cronbach’s α and factor loadings of indicators for each construct support unidimensionality (See Table 5.7). Convergent validity is assessed by performing CFA on a multiple indicator measurement model. Significant t-values of path coefficients in the CFA model provide evidence of convergent validity (Lusch & Brown,

1996) (See Table 5.8).

104

Item Factor % of variance Cronbach’s loading explained α Perceived product quality 60.18 .82 PQ1 .83 PQ2 .79 PQ3 .71 Perceived risk 41.65 .67 PR2 .69 PR3 .53 PR4 .70 Attitude toward a product 80.15 .92 A1 .88 A2 .90 A3 .91 Behavioral intention 55.12 .77 I1 .92 I2 .57 I3 .64

Table 5.7. Results from EFA of the Finalized Measurements

105

Parameter Est. SE t Path coefficients Product quality (ξ1) → PQ1 λ11 .63 .02 31.92*** Product quality (ξ1) → PQ2 λ21 .60 .02 29.01***

Product quality (ξ1) → PQ3 λ31 .94 .01 64.69***

Perceived risk (ξ2) → PR2 λ42 .71 .03 27.41***

Perceived risk (ξ2) → PR3 λ52 .55 .03 20.81***

Perceived risk (ξ2) → PR4 λ62 .67 .03 25.75***

Attitude toward a product (ξ3) → A2 λ73 .88 .01 111.91*** Attitude toward a product (ξ3) → A3 λ83 .89 .01 119.48*** Attitude toward a product (ξ3) → A4 λ93 .92 .01 142.62*** Behavioral Intention (ξ4) → I2 λ10,4 .88 .01 64.54*** Behavioral Intention (ξ4) → I2 λ11,4 .59 .02 28.08*** Behavioral Intention (ξ4) → I3 λ12,4 .72 .02 41.16*** Factor covariances Product quality (ξ1) ↔ Perceived risk (ξ2) φ21 -.05 .04 -1.51 Product quality (ξ1) ↔ Attitude toward a product (ξ3) φ31 .76 .02 44.01*** Product quality (ξ1) ↔ Behavioral Intention (ξ4) φ41 .59 .03 24.23*** Perceived risk (ξ2) ↔ Attitude toward a product (ξ3) φ32 -.18 .03 -5.45*** Perceived risk (ξ2) ↔ Behavioral Intention (ξ4) φ42 -.36 .03 -10.88*** Attitude toward a product (ξ3) ↔ Behavioral Intention (ξ4) φ43 .63 .02 29.92*** Error covariances/variances

δ1 ↔ δ1 Θδ1 .62 .03 24.71*** δ2 ↔ δ2 Θδ2 .65 .03 26.43*** δ3 ↔ δ3 Θδ3 .12 .03 4.57*** δ1 ↔ δ2 Θδ1,2 .28 .02 13.51*** δ4 ↔ δ4 Θδ4 .50 .04 13.69*** δ5 ↔ δ5 Θδ5 .70 .03 23.67*** δ6 ↔ δ6 Θδ6 .56 .03 16.23*** δ7 ↔ δ7 Θδ7 .23 .04 16.49*** δ8 ↔ δ8 Θδ8 .21 .01 15.81*** δ9 ↔ δ9 Θδ9 .16 .01 13.55*** δ10 ↔ δ10 Θδ10 .22 .02 9.00*** δ11 ↔ δ11 Θδ11 .65 .03 26.54*** δ12 ↔ δ12 Θδ12 .49 .03 19.79*** Model fit Chi-square 192.24 (df = 47) RMSEA .049 90 % C.I. (.042; .056) p for the test of perfect fit .000 p for the test of close fit .595 TLI .97 NFI .96 IFI .97 Note. *p<.05, **p<.01, ***p<.001.

Table 5.8. Results from CFA of the Finalized Measurements

106

Discriminant validity refers to “the degree to which measures of different concepts are distinct” (Bagozzi & Phillips, 1991, p. 425). The discriminant validity is achieved when the confidence interval (CI) of the correlation estimate does not contain

1.0 (Anderson & Gerbing, 1988). Since the CIs of the correlation coefficients of the four latent variables (i.e., perceived product quality, perceived risk, attitude toward a product, and behavioral intentions) did not include 1.0, the discriminant validity of the measures in this study was established (Table 5.9). In addition, discriminanat validity can be addressed by performing chi-square difference tests (Anderson & Gerbing, 1988). Chi- square difference tests can be conducted by comparing between an unconstrained model and a constrained model which sets one factor correlation to 1.0. Significant chi-square differences indicate that constrained models are not equivalent to the unconstrained model, implying that the two constructs are not perfectly correlated (Bagozzi, Yi, &

Phillips, 1982). Since chi-square difference tests were significant in this study (See Table

5.10), discriminant validity is achieved.

107

Correlation SE Confidence interval coefficient Product quality ↔ Perceived risk -.05 .04 .03 -.13 Product quality ↔ Attitude toward a .76 .02 .8 .72 product Product quality ↔ Behavioral .59 .03 .65 .53 Intention Perceived risk ↔ Attitude toward a -.18 .03 -.12 -.24 product Perceived risk ↔ Behavioral -.36 .03 -.3 -.42 Intention Attitude toward ↔ Behavioral .63 .02 .67 .59 a product Intention

Table 5.9. Correlation Coefficients and Confidence Interval for Discriminant Validity

Constraint Chi-square df Chi-square df difference difference Unconstrained model 192.24 47 1 Perceived product quality & Perceived risk 786.63 48 594.39*** 1

Perceived product quality & Attitude toward 454.05 48 261.81*** 1 a product Perceived product quality & Behavioral 598.46 48 406.22*** 1 intentions Perceived risk & Attitude toward a product 761.66 48 596.42*** 1 Perceived risk & Behavioral intentions 670.15 48 477.91*** 1 Attitude toward a product & Behavioral 804.15 48 611.91*** 1 intentions Note. *p<.05, **p<.01, ***p<.001.

Table 5.10. Chi-square Difference Test for Discriminant Validity

108

Manipulation Checks

Manipulation checks were performed to examine if participants perceived different experimental conditions as intended. Participants were asked to assess haptic imagery-provoking ability of verbal descriptions and picture information. Two items for verbal information (i.e., ‘verbal descriptions of the dress had features to help me feel fabric properties’ and ‘verbal descriptions of the dress helped me imagine the touch of the fabric’) and two items for pictorial information [i.e., ‘pictures (e.g., dress, fabric swatch) had features to help me feel fabric properties’ and ‘pictures (e.g., dress, fabric swatch) helped me imagine the touch of the fabric’] used 7-point Likert scales ranging from

“strongly agree (7)” to “strongly disagree (1).” Since haptic imagery-provoking items for verbal information (Cronbach’s α =.91) and items for pictorial information (Cronbach’s α

= .86) were reliable, each measure was averaged to test for significant differences in manipulations. In addition, persuasiveness and believability of verbal descriptions were assessed to ensure that the two verbal conditions were not different from each other in terms of persuasiveness and believability of statements. Participants were also asked to evaluate the extent to which the mock website (or mock catalog) was realistic. These three items were measured with 7-point Likert scales ranging from “strongly agree (7)” to

“strongly disagree (1).”

Results from t-tests revealed that there were significant differences between the two verbal haptic imagery descriptions (high haptic imagery descriptions vs. low haptic imagery descriptions) and between the two conditions of pictorial information (picture swatch vs. no picture swatch). There were no significant differences between high imagery descriptions and low imagery descriptions in terms of believability or

109

persuasiveness, indicating that both descriptions were relatively equally believable and

persuasive. Therefore, manipulations of pictorial information and verbal information

were perceived as intended. However, the manipulation checks of shopping contexts

revealed that the mock catalog pages were perceived as less realistic than mock websites.

This might be because the catalog pages were presented on a computer screen instead of

using more realistic paper catalog pages. However, the mean values of realism in both

conditions were greater than the neutral value which is 4.0, implying that both the mock

website and the mock catalog were somewhat realistic (See Table 5.11).

N Mean SD T df P Pictorial information Haptic imagery-provoking Picture swatch 661 5.42 1.22 3.11 1297 .002 No swatch 638 5.20 1.29 Verbal information Haptic imagery-provoking High haptic imagery description 664 5.07 1.29 3.43 1297 .001 Low haptic imagery description 635 4.82 1.28 Believable High haptic imagery description 662 5.31 1.17 .78 1289 .435 Low haptic imagery description 629 5.26 1.16 Persuasive High haptic imagery description 657 4.91 1.33 1.28 1285 .200 Low haptic imagery description 630 4.82 1.26 Shopping context Realistic Online shopping context 633 5.30 1.43 2.01 1293 .046 Catalog shopping context 662 5.14 1.41

Table 5.11. t-test Results for the Manipulation Checks

110

Hypotheses Testing

Hypotheses 1 through 5 examined the effects of pictorial information, verbal information, and shopping context on perceived product quality and perceived risk and the moderating effect of NFT on perceived product quality and perceived risk (Part I).

Hypotheses 6 through 10 addressed the relationships among perceived product quality, perceived risk, attitude toward a product, and behavioral intentions (Part II).

Stimulus Organism Response (Environmental Cues) (Internal States) (Shopping Outcomes)

PART II PART I

Information Presentation Perceived Product Quality Pictorial information

Attitude Behavioral toward Verbal Information Intentions a Product

Shopping context Perceived Risk

Individual characteristics

NFT

Figure 5.2. Proposed Model of the Study 111

Part I (Analysis of Variance)

The first part of the model (H1 trough H5) is tested by using a 2 x 2 x 2 x 2 between-subjects multivariate analysis of variance. Independent variables were pictorial information (picture swatch or no swatch), verbal information (high haptic imagery description or low haptic imagery description), shopping context (online shopping or catalog shopping), and NFT (high NFT or low NFT). Dependent variables were three dimensions of perceived product quality (usefulness, performance quality and aesthetic quality) and three dimensions of perceived risk (financial risk, psychological risk and performance risk).

Hypotheses H1 Pictorial information associated with high haptic imagery will have a positive effect on consumer internal states (a) more positive perceptions of product quality and b) less perceived risk ) compared to pictorial information associated with low haptic imagery. H2 Verbal information associated with high haptic imagery will have a positive effect on consumer internal states (a) more positive perceptions of product quality, b) less perceived risk) compared to verbal information associated with low haptic imagery. H3 Pictorial information and verbal information will interact to affect consumer internal states (a) perceived product quality and b) perceived risk).

H4 The Internet shopping context will have a positive effect on internal states (a) more positive perceptions of product quality and b) less perceived risk) as compared to a catalog shopping context. H5 Need for touch will moderate the relationship between information presentation and consumer internal states (a) perceived product quality and b) perceived risk).

Table 5.12. Summary of H1 through H5

112

Results from the MANOVA revealed that there were no main or interactive

effects for pictorial information and verbal information on perceived product quality and

perceived risk. Therefore, H1, H2 and H3 were not supported. In addition, the results

showed that there were no significant differences between online shopping context and

catalog shopping context in terms of perceived product quality and perceived risk. Thus,

H4 was not supported (See Table 5.13).

Hypothesis 5 postulated that NFT would moderate the effect of product information presentation on customer internal states (perceived product quality and perceived risk) and can be tested by the interaction effects between NFT group and pictorial information and between NFT group and verbal information. There were no interaction effects between NFT and product information presentation. Therefore,

Hypothesis 5 was not supported.

However, this analysis found a significant main effect for NFT on dependent

variables, Wilk’s λ= .95, F (6, 1278) = 12.34, p =.000. The follow up ANOVA results revealed that high NFT subjects perceived the apparel to have higher performance quality

(M = 5.34, SD = 1.04) and higher aesthetic quality (M = 5.59, SD = 1.14) than the low

NFT group (M = 5.13, SD = 1.09 for perceived performance quality; and M = 5.40, SD =

1.14 for perceived aesthetic quality). In addition, there were significant differences in perceived risk between high NFT and low NFT groups. High NFT subjects perceived greater financial risk (M = 5.39, SD = 1.27), psychological risk (M = 5.83, SD = 1.13), and performance risk (M = 4.86, SD = 1.38) than low NFT subjects [financial risk (M =

4.91, SD = 1.28), psychological risk (M = 5.54, SD = 1.12), and performance risk (M =

113

4.43, SD = 1.42)]. Therefore, although high NFT people evaluated apparel quality more positively, they perceived greater risk as compared to low NFT people (See Table 5.14).

Wilk’s Partial Source F df p λ η2 Pictorial information (P) .999 .203 6,1278 .976 .001 Verbal information (V) .996 .886 6, 1278 .504 .004 Shopping context (S) .998 .410 6, 1278 .873 .002 NFT group (N) .945 12.337 6, 1278 .000 .055 P x V .998 .488 6,1278 .818 .002 P x S .995 1.034 6,1278 .401 .005 P x N .996 .78 6, 1278 .586 .004 V x S .998 .507 6, 1278 .803 .002 V x N .993 1.437 6,1278 .197 .007 S x N .995 1.137 6, 1278 .338 .005 P x V x S .996 .888 6, 1278 .502 .004 P x V x N .994 1.344 6, 1278 .234 .006 P x S x N .996 .911 6,1278 .486 .004 V x S x N .992 1.672 6,1278 .124 .008 P x V x S x N .991 1.872 6, 1278 .082 .009 Note. Dependent variable = perceived product quality (usefulness, performance quality and aesthetic quality) and perceived risk (financial risk, psychological risk and performance risk)

Table 5.13. MANOVA Results for Hypotheses Testing (H1 through H5)

114

High NFT group Low NFT group F Partial M SD N M SD N MS p (1,1283) η2 Perceived quality Usefulness 5.08 1.06 650 4.98 1.03 649 3.17 2.89 .089 .002 Performance quality 5.34 1.04 650 5.13 1.09 649 12.83 11.24 .001 .009 Aesthetic quality 5.59 1.14 650 5.40 1.14 649 11.45 8.87 .003 .007 Perceived risk Financial risk 5.39 1.27 650 4.91 1.28 649 75.16 46.21 .000 .035 Psychological risk 5.83 1.13 650 5.54 1.12 649 27.67 21.62 .000 .017 Performance risk 4.86 1.38 650 4.43 1.42 649 56.52 28.95 .000 .022

Table 5.14. ANOVA Results for the Comparisons between High NFT and Low NFT

Groups

Part II (Structural Equation Modeling)

The part II of the proposed model examined the relationships among perceived product quality, perceived risk, attitude toward a product and behavioral intentions by

using structural equation modeling (SEM). Thus, a structural equation model which

consisted of 4 latent variables and 12 manifest variables was constructed to test H6

through H10 and consisted of two exogenous latent variables -- perceived product

quality (ξ1) and perceived risk (ξ2) and two endogenous latent variables -- attitude toward

a product (η1) and behavioral intentions (η2). Using RAMONA (RAM Or Near

Approximation), Maximum Likelihood function was employed to estimate parameters of

the structural equation model and test the relationships among variables. H6 predicted the

115

positive effect of perceived product quality on attitude toward a product (γ1) while H7 addressed the negative effect of perceived risk on attitude toward a product (γ2). In addition, H8 suggested the positive effect of perceived product quality on behavioral intentions (γ3) while H9 postulated the negative effect of perceived risk on behavioral intentions (γ4). Finally, H10 predicted the positive effect of attitude toward a product on behavioral intentions (β1) (See Figure 5.3).

116

Θδ1,2

θδ1 θδ2 θδ3

δ1 δ2 δ3

Θε1 Θε2 Θε3 1 1 1

PQ1 PQ2 PQ3 ε1 ε2 ε3

λx1 λx2 λx3 1 11

A2 A3 A4 Perceived λ λ Product y1 y2 λy3 ψ ζ Quality (ξ1) 2 2 γ 1 1 3 1 1 γ 1 Attitude Behavioral φ 1 toward a Intentions β Product (η1) 1 (η2) γ 1 1 2 ζ1 ψ1

Perceived λy4 λy5 λy6 Risk γ4 (ξ2) I1 I2 I3

1 1 1

λx4 λx5 λx6 ε4 ε5 ε6

PR2 PR3 PR4

Θε4 Θε5 Θε6 1 1 1

δ4 δ5 δ6

Θδ4 Θδ5 Θδ6

Figure 5.3. A Structural Equation Model for Testing H6 through H10

117

Overall Model Fit

Before considering the hypotheses testing, the overall fit of the model was assessed. The chi-square statistic was 192.24 (df = 47, p = .000), indicating that the proposed model failed to fit the data. However, since chi-square test statistics reflect the sensitivity to the large sample size, other fit indices were considered to evaluate the fit of the model (Bagozzi & Yi, 1988). The point estimate of RMSEA was .049 with a confidential interval between .042 and .056, indicating a close and reasonable fit of the hypothesized model to the data. The value of Tucker-Lewis Index (TLI) was .97, Normed

Fit Index (NFI) was .98, and Incremental Fit Index (IFI) was .98. The fit indices (TLI,

NFI, IFI) indicated the good overall model fit since they are greater than .95 (See Table

5.15).

Hypotheses Testing

H6: Perceived product quality will be positively associated with attitude toward a product.

The estimates of the SEM and significant levels of the estimates were presented in the Table 5.15. H6 addresses the positive relationship between perceived product quality and attitude toward a product. The results showed a significant positive effect of perceived product quality on attitude toward a product (γ1 =.76, t =42.62, p =.000).

Therefore, H6 were supported.

118

H7: Perceived risk will be negatively associated with attitude toward a product.

H7 predicted the negative relationships between perceived risk and attitude toward a product. Path coefficients of the SEM found significant negative relationships between perceived risk and attitude toward a product (γ2 = -.14, t = -5.79, p =.000), supporting H7.

H8: Perceived product quality will be positively associated with behavioral intentions.

H8 postulated the positive relationships between perceived product quality and behavioral intentions. The result showed the significant positive relationships between perceived risk and attitude toward a product (γ3 = .311, t = 6.63, p =.000). Therefore H8 was supported.

H9: Perceived risk will be negatively associated with behavioral intentions.

H9 proposed the negative relationships between perceived risk and behavioral intentions. The results demonstrate a significant negative impact of perceived risk on behavioral intentions (γ4 = -.28, t = -9.57, p =.000), supporting H9.

119

H10: Attitude toward a product will positively influence behavioral intentions.

H10 predicted a positive relationship between attitude toward a product and

behavioral intentions. Results revealed a significant positive relationship between attitude

toward a product and behavioral intentions (β1 =.34, t =7.28, p =.000). Therefore, H10

was supported.

The SEM model allowed examination of indirect and direct effects from

perceived quality to behavioral intentions and from perceived risk to behavioral

intentions. MacKinnon, Warsi, and Dwyer (1995) explained methods of calculating direct,

indirect, and total effects among three variables by using mediated effect estimation. One

of the methods can be addressed in this study and involves two regression equations1: In

the model 1, (τ’) refers to the coefficient relating an independent variable to a dependent variable; (β) is the coefficient relating a mediator to a dependent variable; (α) is the coefficient relating the mediator to independent variable. Indirect effect is calculated by

the product of two parameter (αβ), and then total effect becomes (αβ + τ’= τ). Thus, from

1 The two regression models explain the figure below. Model 1 (Y0 = τ’ Xp+ β Xm + ε1) and Model 2 (Xm

= αXp + ε2) where Y0 is the dependent variable, Xp is the independent variable, Xm is the mediator, ε1 and ε2 codes unexplained variability, and the intercept is assumed to be zero.

αβMediator

Independent Dependent variable variable τ’ Indirect effect = αβ; Direct effect = τ’ ; Total effect = αβ + τ’= τ

Figure. A Three Variable Mediation Model (MacKinnon, Warsi, & Dwyer, 1995, p. 45)

120

the SEM model of the study, it is possible to find indirect and direct effects of perceived

product quality and perceived risk on purchase intentions. The total effect of perceived

product quality on behavioral intentions (γ3 + γ1β1) equals the direct effect (γ3) and the indirect effect (γ1β1), and the total effect of perceived risk on behavioral intentions (γ4 +

γ2β1) consists of the direct effect (γ4) and the indirect effect (γ2β1). Results revealed that

perceived product quality influenced behavioral intentions directly (γ3 =.31) and

indirectly (γ1β1 =.26) through attitude toward a product. Results also showed that the

direct effect of perceived risk on behavioral intentions (γ4= -.28) was stronger than the

indirect effect of perceived risk on purchase intentions (γ2β1 = -.05) through attitude

toward a product (See Table 5.16).

121

Parameter Est. SE t Structural path H6 Product quality (ξ1) → Attitude toward a product (η1) γ1 .76 .02 42.62*** H7 Perceived risk (ξ2) → Attitude toward a product (η1) γ2 -.14 .03 -5.79*** H8 Product quality (ξ1) → Behavioral Intention (η2) γ3 .31 .05 6.63*** H9 Perceived risk (ξ2) → Behavioral Intention (η2) γ4 -.28 .03 -9.57*** H10Attitude toward a product (η1) → Behavioral Intention (η2) β1 .34 .05 7.28*** Measurement model Product quality (ξ1) → PQ1 λx1 .63 .02 31.92*** Product quality (ξ1) → PQ2 λx2 .60 .02 29.01*** Product quality (ξ1) → PQ3 λx3 .94 .01 64.69*** Perceived risk (ξ2) → PR2 λx4 .71 .03 27.41*** Perceived risk (ξ2) → PR3 λx5 .55 .03 20.81*** Perceived risk (ξ2) → PR4 λx6 .67 .03 25.75*** Attitude toward a product (η1) → A2 λy1 .88 .01 111.91*** Attitude toward a product (η1) → A3 λy2 .89 .01 119.48***

Attitude toward a product (η1) → A4 λy3 .92 .01 142.63*** Behavioral Intention (η2) → I1 λy4 .88 .01 64.54*** Behavioral Intention (η2) → I2 λy5 .59 .02 28.08*** Behavioral Intention (η2) → I3 λy6 .72 .02 41.46*** Factor covariance Product quality (ξ1) ↔ Perceived risk (ξ2) φ1 -.04 .04 -1.51 Error variances/covariances ζ1 ↔ ζ1 Ψ1 .40 .03 15.24*** ζ2 ↔ ζ2 Ψ2 .50 .03 18.50*** δ1 ↔ δ1 Θδ1 .61 .03 24.71*** δ2 ↔ δ2 Θδ2 .65 .02 26.43*** δ1 ↔ δ2 Θδ12 .28 .02 13.51*** δ3 ↔ δ3 Θδ3 .12 .03 4.57*** δ4 ↔ δ4 Θδ4 .50 .04 13.69*** δ5 ↔ δ5 Θδ5 .70 .03 23.67*** δ6 ↔ δ6 Θδ6 .56 .03 16.23*** ε1 ↔ ε1 Θε1 .23 .01 16.49***

ε2 ↔ ε2 Θε2 .21 .01 15.81***

ε3 ↔ ε3 Θε3 .16 .01 13.55*** ε4 ↔ ε4 Θε4 .22 .02 9.00***

ε5 ↔ ε5 Θε5 .65 .03 26.54***

ε6 ↔ ε6 Θε6 .49 .03 19.79*** Model fit Chi-square 192.24 (df = 47) RMSEA .049 90 % C.I. (.042; .056) p for the test of perfect fit .000 p for the test of close fit .595 TLI .97 NFI .98 IFI .98 Note. *p<.05, **p<.01, ***p<.001.

Table 5.15. Results from the SEM for Testing H6 through H10 122

Indirect Direct Total Dependent variables Predictor variables effects effects effects Behavioral intentions Perceived quality .26 .31 .57 Perceived risk -.05 -.28 -.32 Attitude toward a product

Table 5.16. Decomposition of Direct, Indirect, and Total Effects for the Hypothesized

Model.

Post Hoc Analysis (An Alternative Model)

This section presents an alternative model since the Part I of the proposed model

was not supported. The study failed to find evidence of the effect of the pictorial and

verbal information and shopping contexts on customer cognitive evaluations (i.e.,

perceived product quality and perceived risk), and the moderating role of NFT on

influencing the effect of product presentation on customer internal states.

Thus, the alternative model postulated the indirect effect of pictorial and verbal information and shopping contexts on perceived product quality and perceived risk. The model predicted that certain product information (e.g., fabric swatch presentation, fabric hand descriptions) stimulates haptic imagery process and then highly evoked haptic imagery influences positive perceptions of product quality and reduce perceived risk. In addition, the model suggests that the online shopping context conveys greater perceived interactivity as compared to the catalog shopping context, and high perceived

123

interactivity influences positive perceptions of product quality and decreases shoppers’

perceived risk. Therefore, according to the alternative model, new hypotheses (H11 through H16) were developed. In addition, H6 through H10 from the original model were tested in the alternative model (See Figure 5.4).

H11: Pictorial information associated with a picture swatch will evoke greater haptic imagery compared to pictorial information not associated with a picture swatch.

H12: Verbal information associated with fabric hand descriptions will evoke greater haptic imagery compared to verbal information associated with style descriptions.

H13: Need for touch will moderate the effect of pictorial and verbal information on haptic imagery.

H14: The Internet shopping context will be perceived as more interactive as compared to a catalog shopping context.

H15: Perceived haptic imagery will positively influence customers’ internal states (a) positively influence perceptions of product quality and b) negatively influence perceived risk).

H16. Perceived interactivity will positively influence customers’ internal states (a) positively influence perceptions of product quality and b) negatively influence perceived risk).

124

Stimulus Organism Response

(Environmental Cues) (Internal States) (Shopping Outcomes)

PART I PART II

Information presentation Haptic Perceived Pictorial information Imagery Product Quality

Attitude Behavioral Verbal Information toward Intentions a Product

Interactivity Perceived Shopping context Risk

Individual characteristics

NFT

Figure 5.4. The Alternative Model of the Study

Descriptive and Factor Analyses of Additional Measurements

The alternative model of the study added two more variables (perceived haptic

imagery and perceived interactivity) to the original model. Thus, the section of the study

125

presents the descriptive statistics of the additional variables (See Table 5.18) and exploratory factor analyses of one variable (haptic imagery).

Haptic imagery

Fourteen-items were developed in order to measure perceived haptic imagery.

Among 14 items, 7 items were adopted from Ellen and Bone (1991)’s Communication-

Evoked Imagery Processing scale and revised for this study. The other 7 items addressing specific haptic properties (i.e., texture, lightness, softness, smoothness, coolness,

drapability, and silkiness) were developed. In order to identify possible dimensions of

perceived haptic imagery, Maximum Likelihood (ML) model estimation and oblique

rotation with CF-Quartimax criterion were employed by using the CEFA 2.0. While examining 2, 3, and 4 factor solutions, conceptual content and measures of model fit were considered. In addition, each factor model was re-examined after deleting items with low factor loadings (lower than .40). Based on Kaiser’s Criterion (1960) and the Scree test

(Cattell, 1966), a three factor model was recommended because three eigenvalues were greater than 1.0. In addition, a three factor model was acceptable in terms of content

considerations of constructs.

The final EFA provided a three factor model with 8 items (See Table 5.17). The

RMSEA estimate indicated a close fit (.044) for the model. The first factor was labeled as

‘vividness’ and referred to the extent of perceived imagery vividness. Three items were

used for the first factor (i.e., vivid, detailed, unclear). The second factor was labeled as

‘imagery elaboration’ and addressed “the activation of stored information in the

production of mental images beyond what was provided by the stimulus” (Babin & Burns,

126

1997, p. 37). The second factor was measured with two items (e.g., I fantasized what it would be like to touch the dress). The last factor was labeled as perceived ‘imagery of haptic

properties’ and referred to mental imagination of fabric properties which are generally

achieved by a sense of touch. The items of the third factor were associated with three

pleasant haptic properties (i.e., smoothness, softness, and silkiness). Scores on the items

within each factor were averaged and used as indicators of the latent variable of haptic

imagery. Descriptive statistics for the scale was presented in Table 5.18. Cronbach’s α for

vividness was .73 (α for online = .76, α for catalog = .70), Cronbach’s α for imagery elaboration was .76 (α for online = .76, α for catalog = .75), and Cronbach’s α for imagery of haptic properties was = .79 (α for online = .79, α for catalog = .79) (See Table

5.18).

127

Factor loadings Items 1 2 3

Factor 1: Vividness HI13 The imagery which occurred was detailed. .756 HI11 The imagery which occurred was vivid. .646 HI12* The imagery which occurred was unclear. .577 Factor 2: Imagery Elaboration HI9 I fantasized what it would be like to touch the dress. 1.10 HI8 I imagined the feel of fabric textures of the dress. .48 Factor 3: Imagery of Haptic Properties HI4 When I browse for the information about the dress in .829 the website (catalog), I got a sense of smoothness of the fabric. HI3 When I browse for the information about the dress in .771 the website (catalog), I got a sense of softness of the fabric. HI7 When I browse for the information about the dress in .540 the website (catalog), I got a sense of silkiness of the fabric. Eigenvalue 3.51 1.21 1.1 Sample discrepancy function value .018 χ2(df) 25.32 (7) p for the test of exact fit .001 p for the test of close fit .684 RMSEA (C.I.) .044 (.026; .063) Note. *item was reversed.

Table 5.17. EFA Results for the 8item, 3-Factor Model for Haptic Imagery

128

Perceived Interactivity

Perceived interactivity was measured by using 5 items developed by Schlosser

(2003). The perceived interactivity scale was found to be reliable in this study

(Cronbach’s α for the overall sample = .86, α for online = .87 and α for catalog = .85).

Descriptive statistics of 5 items were presented in Table 5.18.

Total sample Online Catalog

(N = 1299) (N = 635) (N = 664) Cronbach’s Cronbach’s Cronbach’s Min. Max. M SD α M SD α M SD α

Haptic imagery H1 Vividness 1 7 4.43 4.51 .73 4.81 1.14 .76 4.42 1.48 .70 H2 Imagery elaboration 1 7 4.43 1.50 .76 4.44 1.54 .76 4.76 1.12 .75 H3 Imagery of haptic 1 7 4.44 1.24 .79 4.42 1.25 .79 4.46 1.22 .79 properties Interactivity IN1 1 7 5.37 1.27 .86 5.40 1.26 .87 5.35 1.27 .85 IN2 1 7 5.05 1.44 5.08 1.46 5.02 1.42 IN3 1 7 4.82 1.35 5.00 1.30 4.65 1.37 IN4 1 7 4.80 1.35 4.95 1.33 4.66 1.36 IN5 1 7 4.81 1.37 4.99 1.21 4.630 1.40

Table 5.18. Descriptive Statistics and Reliability for the Additional Measures

129

Initial Model Considerations

Confirmatory factor analysis (CFA) was used to assess convergent and discriminant validity of the measurements and exploratory factor analysis (EFA) was used to address unidimensionality. Using RAMONA, Maximum Likelihood function was employed to conduct CFA. The alternative measurement model was respecified based on

Anderson and Gerbing’s (1988) two-step modeling approach. Respecification was taken into consideration if an indicator contained a low factor loading (e.g., lower than .60) and a low squared multiple correlation values (e.g., lower than .40), as compared to other indicators from the same factor (Bagozzi & Yi, 1988).

IN1 and IN2 (the first and second indicators for perceived interactivity) were deleted because they had low factor loadings (.40 and .55) and low SMCs (.16 and .30) as compared to other indicators of interactivity (IN4, IN5, and IN6). Therefore, the measurement model for the alternatively proposed study was finalized, and the results of the finalized CFA are presented in Figure 5.5 and Table 5.20.

130

Θδ7.8

θδ1 θδ2 θδ3 Θδ4 Θδ5 Θδ6 Θδ7 Θδ8 Θδ9

δ1 δ2 δ3 δ4 δ5 δ6 δ7 δ8 δ9

1 1 1 1 11 1 1 1

H1 H2 H3 IN3 IN4 IN5 PQ1 PQ2 PQ3

λ11 λ21 λ31 λ λ λ 42 52 λ 83 λ 62 λ73 93

Haptic Perceived Imagery Interactivity Product (ξ2) Quality (ξ3) (ξ1)

1 1 1

Φ21. . . Φ65

1 1 1

Behavioral Perceived Attitude Intentions Risk (ξ4) (ξ5) (ξ6)

λ18,6 λ λ13,5 λ16,6 λ10,4 12,4 λ λ 14,5 15,5 λ17,6 λ11,4 PR1 PR2 PR3 A2 A3 A4 I1 I2 I3

1 1 1 1 1 1 1 1 1

δ10 δ11 δ12 δ13 δ14 δ15 δ16 δ17 δ18

Θδ10 Θδ11 Θδ12 Θδ13 Θδ14 Θδ15 Θδ16 Θδ17 Θδ18

Figure 5.5. A CFA Model of Final Measurement Items for the Alternative Model

131

The overall fit of the measurement model was assessed. Although the chi-square statistic was significant (χ2 = 456.89, df = 119, p = .000), indicating that the proposed model failed to fit the data, other fit indices were considered because chi-square statistic tends to have sensitivity to large sample size (N=1299). The point estimate of RMSEA was .047 with a confidential interval between .042 and .051, indicating a close fit of the model to the data. The other fit indices (TLI = .96, NFI =.96, and IFI =.97) supported the good overall model fit.

Unidimensionality was assessed by an exploratory analysis for each construct.

Each exploratory factor analysis produced one factor solution and Cronbach’s α provided reliability for each construct. Thus, unidimentianlity was established (See Table 5.19).

Convergent validity was assessed by using a CFA of a multiple indicator measurement model. The CFA results supported unidimensionality since each latent construct consisted of multiple measures and each measure belonged to only one latent construct

(Anderson, Gerbing, & Hunter, 1987). The CFA results also revealed that t-values of path coefficients were significant, supporting the convergent validity of the measures (Lusch

& Brown, 1996) (See table 5.20).

132

Item Factor loading % of variance explained Cronbach’s α Haptic imagery 42.07 .68 H1 .62 H2 .68 H3 .65 Interactivity 75.03 .89 IN1 .80 IN2 .95 IN3 .84

Table 5.19. Results from EFA of the Added Measurements for the Alternative Model

133

Para. Est. SE t Path coefficients Haptic imagery (ξ1) → H1 λ11 .68 .02 28.46*** Haptic imagery (ξ1) → H2 λ21 .64 .02 26.47*** Haptic imagery (ξ1) → H3 λ31 .62 .03 25.16*** Interactivity (ξ2) → IN3 λ42 .81 .01 71.78*** Interactivity (ξ2) → IN4 λ52 .93 .01 120.35*** Interactivity (ξ2) → IN5 λ62 .86 .01 88.38*** Product quality (ξ3) → PQ1 λ73 .65 .02 34.98*** Product quality (ξ3) → PQ2 λ83 .62 .02 31.81*** Product quality (ξ3) → PQ3 λ93 .90 .01 67.90*** Perceived risk (ξ4) → PR2 λ10,4 .70 .03 27.40*** Perceived risk (ξ4) → PR3 λ11,4 .56 .03 20.95*** Perceived risk (ξ4) → PR4 λ12,4 .67 .03 25.99*** Attitude toward a product (ξ5) → A2 λ13,5 .88 .01 111.90*** Attitude toward a product (ξ5) → A3 λ14,5 .89 .01 120.14*** Attitude toward a product (ξ5) → A4 λ15,5 .92 .01 141.44*** Behavioral Intention (ξ6) → I2 λ16,6 .87 .01 64.68*** Behavioral Intention (ξ6) → I2 λ17,6 .58 .02 27.76*** Behavioral Intention (ξ6) → I3 λ18,6 .79 .02 44.00*** Factor covariances Haptic imagery (ξ1) ↔ Interactivity (ξ2) φ21 .44 .03 14.77*** Haptic imagery (ξ1) ↔ Product quality (ξ3) φ31 .47 .03 14.99*** Haptic imagery (ξ1) ↔ Perceived risk (ξ4) φ41 -.13 .04 -3.19*** Haptic imagery (ξ1) ↔ Attitude toward a product (ξ5) φ51 .36 .03 11.27*** Haptic imagery (ξ1) ↔ Behavioral Intention (ξ6) φ61 .41 .03 12.36*** Interactivity (ξ2) ↔ Product quality (ξ3) φ32 .46 .03 17.67*** Interactivity (ξ2) ↔ Perceived risk (ξ4) φ42 -.19 .03 -5.71*** Interactivity (ξ2) ↔ Attitude toward a product (ξ5) φ52 .36 .03 13.33*** Interactivity (ξ2) ↔ Behavioral Intention (ξ6) φ62 .45 .03 17.20*** Product quality (ξ3) ↔ Perceived risk (ξ4) φ43 -.06 .04 -1.72 Product quality (ξ3) ↔ Attitude toward a product (ξ5) φ53 .78 .02 47.32***

Product quality (ξ3) ↔ Behavioral Intention (ξ6) φ63 .62 .02 25.64*** Perceived risk (ξ4) ↔ Attitude toward a product (ξ5) φ54 -.19 .03 -5.46***

Perceived risk (ξ4) ↔ Behavioral Intention (ξ6) φ64 -.36 .03 -10.91*** Attitude toward a product (ξ5) ↔ Behavioral Intention (ξ6) φ65 .64 .02 30.28***

Continued

Table 5.20. Results from CFA of the Finalized Measurements for the Alternative Model

134

Table 5.20 continued

Error covariances/variances

δ1 ↔ δ1 Θδ1 .54 .03 16.72*** δ2 ↔ δ2 Θδ2 .59 .03 18.86*** δ3 ↔ δ3 Θδ3 .62 .03 20.26*** δ4 ↔ δ4 Θδ4 .34 .02 18.65*** δ5 ↔ δ5 Θδ5 .14 .01 9.57*** δ6 ↔ δ6 Θδ6 .27 .02 15.94*** δ7 ↔ δ7 Θδ7 .57 .02 23.55*** δ8 ↔ δ8 Θδ8 .61 .02 24.94*** δ7 ↔ δ8 Θδ7,8 .24 .02 12.29*** δ9 ↔ δ9 Θδ9 .20 .02 8.34*** δ10 ↔ δ10 Θδ10 .51 .04 14.03*** δ11 ↔ δ11 Θδ11 .69 .03 23.57*** δ12 ↔ δ12 Θδ12 .55 .03 16.14*** δ13 ↔ δ13 Θδ13 .23 .01 16.46*** δ14 ↔ δ14 Θδ14 .21 .01 15.72*** δ15 ↔ δ15 Θδ15 .16 .01 13.62*** δ16 ↔ δ16 Θδ16 .24 .02 10.41*** δ17 ↔ δ17 Θδ17 .66 .03 26.77*** δ18 ↔ δ18 Θδ18 .47 .02 19.16*** Model fit Chi-square 456. (df = 119) 89 RMSEA .047 90 % C.I. (.042; .051) p for the test of perfect fit .000 p for the test of close fit .875 TLI .96 NFI .96 IFI .97 Note. *p<.05, **p<.01, ***p<.001.

Discriminant validity is also assessed for the measures of the alternative model.

Since the correlation coefficients of four latent variables (i.e., haptic imagery, interactivity, perceived product quality, perceived risk, attitude toward a product, and behavioral intentions) did not contain 1.0 (see Table 5.21), discriminant validity of the measures for the alternative model was achieved. Chi-square difference tests were performed by comparing between an unconstrained model and a constrained model (i.e.,

135

one of factor correlations is fixed to 1.0) (Anderson & Gerbing, 1988). The chi-square difference tests were significant, indicating that discriminant validity was supported (see

Table 5.22).

Correlation SE Confidence interval coefficient Haptic imagery ↔ Interactivity 0.44 0.03 0.38 0.50 Haptic imagery ↔ Product quality 0.47 0.03 0.41 0.53 Haptic imagery ↔ Perceived risk -0.13 0.04 -0.21 -0.05 Haptic imagery ↔ Attitude toward a 0.36 0.03 0.3 0.42 product Haptic imagery ↔ Behavioral 0.41 0.03 0.35 0.47 intention Interactivity ↔ Product quality 0.46 0.03 0.4 0.52 Interactivity ↔ Perceived risk -0.19 0.03 -0.25 -0.13 Interactivity ↔ Attitude toward a 0.36 0.03 0.3 0.42 product Interactivity ↔ Behavioral 0.45 0.03 0.39 0.51 Intention Product quality ↔ Perceived risk -0.06 0.04 -0.14 0.02 Product quality ↔ Attitude toward a 0.78 0.02 0.74 0.82 product Product quality ↔ Behavioral 0.62 0.02 0.58 0.66 Intention Perceived risk ↔ Attitude toward a -0.19 0.03 -0.25 -0.13 product Perceived risk ↔ Behavioral -0.36 0.03 -0.42 -0.3 Intention Attitude toward ↔ Behavioral 0.64 0.02 0.6 0.68 a product Intention

Table 5.21. Correlation Coefficients and Confidence Interval for Discriminant Validity

136

Constraint Chi-square df Chi-square df difference difference Unconstrained model 456.89 119 Haptic imagery & Interactivity 912.26 120 455.37*** 1 Haptic imagery & Perceived product quality 852.66 120 395.77*** 1 Haptic imagery & Perceived risk 1062.98 120 606.09*** 1 Haptic imagery & Attitude toward a product 968.23 120 511.34*** 1 Haptic imagery & Behavioral intention 918.62 120 461.73*** 1 Interactivity & Perceived product quality 970.88 120 513.99*** 1 Interactivity & Perceived risk 1023.00 120 566.11*** 1 Interactivity & Attitude toward a product 2610.85 120 2153.96*** 1 Interactivity & Behavioral intention 1284.32 120 827.43*** 1 Perceived product quality & Perceived risk 1052.88 120 595.99*** 1 Perceived product quality & Attitude toward a 691.72 120 234.83*** 1 product Perceived product quality & Behavioral 825.16 120 368.27*** 1 intentions Perceived risk & Attitude toward a product 1025.83 120 568.94*** 1 Perceived risk & Behavioral intentions 931.93 120 475.04*** 1 Attitude toward a product & Behavioral 1058.29 120 601.40*** 1 intentions Note. *p<.05, **p<.01, ***p<.001.

Table 5.22. Chi-square Difference Test for Discriminant Validity

Alternative Model Testing

Part I (Analysis of Variance)

The first part of the alternative model addressed the influence of pictorial and verbal information on haptic imagery and the effect of shopping context on perceived interactivity, and investigated the moderating effect of NFT on the relationships between product information (pictorial and verbal information) and haptic imagery. A 2 x 2 x 2 x

2 between-subjects multivariate analysis of variance was used to test H11 through H14.

137

Independent variables for the alternative model were the same independent variables of

the originally hypothesized model: pictorial information (picture swatch or no swatch), verbal information (fabric hand descriptions or style descriptions), shopping context

(online shopping or catalog shopping), and NFT (high NFT or low NFT). Fabric hand description condition referred to descriptions which contained information about apparel

and fabric texture (e.g., silkiness, smoothness), pressure (e.g., softness, flexibility), weight (e.g., lightness), and temperature (e.g., coolness). The style description condition did not include fabric hand information but rather information about style details (e.g., natural and slight A-line shape, full skirts are classic feminine shape). Dependent variables were perceived haptic imagery and perceived interactivity. Perceived haptic imagery contained three indicators (i.e., vividness, imagery elaboration, and imagery of haptic properties).

Results from the multivariate analysis of variance are displayed in Table 5.23. The results revealed main effects for pictorial information [Wilk’s λ = .99, F (4, 1280) = 3.89, p = .000, partial η2 =.012] , verbal information [Wilk’s λ = .96, F (4, 1280) = 11.90, p

= .000, partial η2 =.0136], shopping context [Wilk’s λ = .98, F (4, 1280) = 7.35, p = .000,

partial η2 =.022] and NFT [Wilk’s λ = .95, F (4, 1280) = 17.52, p = .000, partial η2 =.052] on the dependent variables. In addition, results revealed interaction effects between pictorial information and shopping context [Wilk’s λ = .99, F (4, 1280) = 2.96, p = .019,

partial η2 =.009] and between shopping context and NFT [Wilk’s λ = .99, F (4, 1280) =

2.50, p = .041, partial η2 =.008] (See Table 5.23).

138

Wilk’s F Partial Source p λ (4,1280) η2 Pictorial information (P)* .988 3.885 .000 .012 Verbal information (V)* .964 11.898 .000 .036 Shopping context (S)* .978 7.354 .000 .022 NFT group (N)* .948 17.521 .000 .052 P x V .996 1.201 .309 .004 P x S* .991 2.962 .019 .009 P x N .997 1.082 .364 .003 V x S .995 1.536 .189 .005 V x N .999 .343 .849 .001 S x N* .992 2.499 .041 .008 P x V x S .998 .749 .558 .002 P x V x N .994 1.991 .093 .006 P x S x N .996 1.435 .220 .004 V x S x N .998 .547 .701 .002 P x V x S x N .996 1.162 .326 .004 Note. *significant sources

Table 5.23. MANOVA Results for the Part I of the Alternative Model.

H11: Pictorial information associated with a picture swatch will evoke greater haptic imagery compared to pictorial information associated without a picture swatch. .

H11 predicted the main effect of pictorial information on haptic imagery. ANOVA results revealed that there was a significant effect for pictorial information on imagery elaboration [F (1, 1283) = 7.72, p = .006, partial η2 =.006] and vividness [F (1, 1283) =

4.96, p = .047, partial η2 =.003]. When a picture swatch was presented (M = 4.84, SD =

1.14), participants had more vivid imagery than when a picture swatch was not presented 139

(M = 4.72, SD = 1.12). In addition, participants who were exposed to a fabric swatch

scored higher on perceived imagery elaboration (M = 4.54, SD = 1.48) as compared to

those who were not exposed to a fabric swatch (M = 4.31, SD = 1.52) (see Table 5.24).

However, there was no significant difference in imagery of haptic properties, whether the

picture swatch was provided or not. Therefore, H11 was partially supported.

H12: Verbal information associated with fabric hand description will evoke greater perceived haptic imagery than verbal information associated with style descriptions.

H12 proposed the main effect for verbal information on perceived haptic imagery.

ANOVA results revealed that verbal information significantly influenced perceptions of

vividness [F (1, 1283) = 4.84, p = .028, partial η2 =.004], perceived imagery elaboration

[F (1, 1283) = 12.88, p = .000, partial η2 =.010], and perceived imagery of haptic

properties [F (1, 1283) = 39.40, p = .000, partial η2 =.030]. When more fabric hand

information was provided (M = 4.85, SD = 1.12), participants reported experiencing more

vivid imagery than when style descriptions were presented (M = 4.71, SD = 1.14). When fabric hand descriptions were available (M = 4.57, SD = 1.49), participants scored higher on imagery elaboration than when style descriptions were available (M = 4.28, SD =

1.50). In addition, participants who were exposed to more descriptive fabric hand information (M = 4.65, SD = 1.25) experienced greater imagery of haptic properties as compared to those who were exposed to style descriptions (M = 4.22, SD = 1.19) (See

Table 5.24). Therefore, H12 was supported.

140

F Partial Source M SD N MS p (1,1283) η2 Pictorial information Vividness Fabric swatch 4.84 1.13 661 4.96 3.95 .047 .003 No swatch 4.72 1.12 638

Imagery elaboration Fabric swatch 4.54 1.48 661 16.47 7.72 .006 .006 No swatch 4.31 1.52 638 Imagery of haptic properties Fabric swatch 4.44 1.26 661 .02 .01 .916 .000 No swatch 4.44 1.21 638

Verbal information Vividness Fabric hand description 4.85 1.12 664 6.08 4.84 .028 .004 Style description 4.71 1.14 635 Imagery elaboration Fabric hand description 4.57 1.49 664 27.49 12.88 .000 .010 Style description 4.28 1.50 635 Imagery of haptic properties Fabric hand description 4.65 1.25 664 58.45 39.40 .000 .030 Style description 4.22 1.19 635

Table 5.24. ANOVA Results for the Effect of Pictorial and Verbal Information on Haptic

Imagery.

H13: Need for touch will moderate the effect of pictorial and verbal information on perceived haptic imagery.

H13 proposed that NFT will influence the effect of product information on haptic imagery, predicting the interaction effect of NFT and pictorial information and the interaction effect of NFT and verbal information. However, there were no significant

141

interaction effects between NFT and pictorial or verbal information. Therefore, H13 was

not supported. However, the results found a main effect for NFT. High NFT people

scored higher on perceived imagery elaboration, perceived vividness, perceived imagery

of haptic properties, and perceived interactivity as compared to low NFT people. Means

and standard deviations for each group on the dependent variables are presented in Table

5.25.

High NFT Low NFT

N= 664 N = 635 Partial Dependent variable M SD M SD MS F p (1,1283) η2 Vividness 4.87 1.16 4.69 1.09 9.12 7.26 .007 .006 Imagery elaboration 4.75 1.48 4.11 1.46 135.3 63.46 .000 .047 9 Imagery of haptic properties 4.53 1.26 4.36 1.22 8.86 5.97 .015 .005 Interactivity 4.94 1.26 4.68 1.20 20.68 13.99 .000 .011

Table 5.25. ANOVA Results for the Effect of NFT on Perceived Haptic Imagery and

Perceived Interactivity.

H14: The Internet shopping context will be perceived as more interactive as compared to a catalog shopping context.

The ANOVA results revealed a significant main effect for shopping context on

interactivity. Participants perceived greater interactivity in the online shopping context

142

(M = 4.98, SD = 1.21) as compared to the catalog shopping context (M = 4.64, SD =

1.24), [MS = 37.47, F (1, 1283) = 25.36, p = .000, partial η2 =.019]. Therefore, H14 was

supported.

MANOVA results revealed interaction effects between shopping context and

pictorial information and between shopping context and NFT although they were not

predicted. When websites provided a picture swatch (M=5.14, SD = 1.18), participants

perceived more interactivity as compared to when websites did not provide a picture swatch (M = 4.82, SD = 1.22). This is likely because the website with a picture swatch provided a function which allowed shoppers to click on an icon so that they can have a

closer view of the fabric. Thus, the “click on” function is likely to have increased

perceived interactivity. However, in the catalog shopping context, perceived interactivity

did not differ in terms of whether a picture swatch was presented (M = 4.67, SD = 1.24)

or not (M = 4.61, SD = 1.24), [F (1, 1283) = 4.26, p = .039, partial η2 =.003] (see Figure

5.6). As compared to websites, the catalog shopping context cannot provide any

interactive functions (e.g., closer view, alternative view, and hyperlink). Thus, there were

no differences in perceived interactivity between catalogs with a picture swatch and

without a picture swatch.

High NFT shoppers engaged in more vivid imagery when they browsed online

(M=4.95, SD = 1.15) as compared to when they browsed in the catalog (M = 4.78, SD =

1.16). Low NFT shoppers experienced a bit more vivid imagery in the catalog shopping

context (M = 4.73, SD = 1.07) as compared to the online shopping context (M = 4.66, SD

= 1.11), F (1, 1283) = 3.91, p = .048, partial η2 =.003]. But, overall the low NFT group reported engaging in less vivid imagery than the high NFT group (see Figure 5.7).

143

5.2 5.14 5.1 5 4.9 4.8 4.82 catalog 4.7 4.67 online 4.6 4.61 Interactivity 4.5 4.4 4.3 No swatch Picture swatch

Figure 5.6. Pictorial Information by Shopping Context on Interactivity

5 4.95 4.9

4.8 4.78 Low NFT 4.73 4.7 High NFT vividness 4.66 4.6

4.5 Catalog Online

Figure 5.7. Shopping Context by NFT on Vividness

144

Part II (Structural Equation Modeling)

The second part of the alternative model addressed the relationships among

perceived haptic imagery, perceived interactivity, perceived product quality, perceived

risk, attitude toward a product, and behavioral intentions, and was tested by using

structural equation modeling (SEM). The structural equation model for the alternative

model contained 6 latent variables and 18 manifest variables. The latent variables

consisted of two exogenous latent variables, which were haptic imagery (ξ1) and

interactivity (ξ2), and four endogenous latent variables, which were perceived product

quality (η1), perceived risk (η2), attitude toward a product (η3), and behavioral intentions

(η4). RAMONA with Maximum Likelihood function was used to test the second part of

the alternative model.

The SEM for the alternative model was used to examine the positive effect of

haptic imagery on perceived product quality (γ1) and the negative effect of haptic

imagery on perceived risk (γ2), and to investigate the positive effect of interactivity on

perceived product quality (γ3) and the negative effect of interactivity on perceived risk

(γ4) (H15 through H16). The remaining part of the model examined the same

relationships with the part II of the originally proposed model in this study (H6 through

H10). Thus, the positive relationships between perceived product quality and attitude

toward a product (β1), between perceived product quality and behavioral intentions (β3), and between attitude toward a product and behavioral intentions (β5) were predicted in

the alternative model. The model also postulated the negative relationships between

perceived risk and attitude toward a product (β2) and between perceived risk and

behavioral intentions (β4) (see Figure 5.8). 145

Θε1,2

θδ1 θδ2 θδ3 Θε1 Θε2 Θε3

Θ Θ δ1 δ2 δ3 ε1 ε2 ε3 ε7 Θε8 ε9

1 1 1 1 11

H1 H2 H3 PQ1 PQ2 PQ3 ε7 ε8 ε9

λx1 λx2 λx3 λy λy2 λy3 1 11

1 A2 A3 A4 Perceived γ Perceived 1 λ λ λ Haptic Product y y8 y9 ψ ζ4 Imagery (ξ1) Quality (η1) 4 β3 1 1 1 γ2 1 β1 1 ζ1 Attitudes Behavioral φ1 ψ1 toward a Intentions β Product (η3) 5 1 (η4) β2 1 1 γ3 ζ3

ψ3 Perceived Perceived λy1 λy1 λy12 β Interactivity Risk 4 γ4 (ξ2) (η2) 1 I1 I2 I3

ζ2 ψ2 1 1 1

λx4 λx5 λx6 λy λy5 λy6 ε10 ε11 ε12

IN1 IN2 IN3 PR2 PR3 PR4

Θ Θ Θ 1 1 1 1 1 1 ε10 ε11 ε12

δ4 δ5 δ6 ε4 ε5 ε6

Θδ4 Θδ5 Θδ6 Θε4 Θε5 Θε6

Figure 5.8. A Structural Equation Model for Testing the Second Part of Alternative

Model

146

Overall Model Fit

Fit indices suggested a good fit of the alternative model. The point estimate of

RMSEA was .048 and the confidential interval was between .044 and .053, which

indicated a close and reasonable fit of the proposed alternative model to the data. In addition, TLI (.96), NFI (.96) and IFI (.97) were greater than .95 and these fit indices indicated the good fit of the alternative model. Although the significant chi-square statistic implied that the alternative model failed to fit the data in an absolute sense (χ2 =

465.81, df = 123, p = .000), it is sensitive to the large sample size (Bagozzi & Yi, 1988).

Thus, the model fit was considered to be fairly good based on other fit indices (e.g.,

RMSEA, TLI) (See Table 5.26).

Hypotheses Testing

H15: Perceived haptic imagery will positively influence customers’ internal states (a) positively influence perceptions of product quality and b) negatively influence perceived risk).

H15-a postulated the positive influence of perceived haptic imagery on customers’ perceptions of product quality and H15-b proposed the negative impact of

perceived haptic imagery on perceived risk. The results for H15-a showed the significant

positive relationship between two constructs (γ1 = .33, t = 9.23, p =.000), implying that

the higher the perceived haptic imagery evoked, the more positively product quality is perceived. Thus, H15-a was supported. However, there was no significant relationship

147

between perceived haptic imagery and perceived risk (γ2 = -.05, t = -1.07). Therefore,

H15-b was not supported (See Table 5.26).

H16. Perceived interactivity will positively influence customers’ internal states (a) positively influence perceptions of product quality and b) negatively influence perceived risk).

H16-a predicted the positive effect of perceived interactivity on customers’

perceptions of product quality and H16-b proposed the negative impact of perceived

interactivity imagery on perceived risk. The results revealed a significant positive

relationship between perceived interactivity and perceptions of product quality (γ3 = .32, t

= 10.14, p =.000) and a significant negative relationship between perceived interactivity and perceived risk (γ4 = -.19, t = -4.73, p =.000). Therefore, H16 was supported.

The remaining part of the alternative model proposed the same hypotheses which

were addressed by H6 through H10 in the original model. Path coefficients of the SEM

found that perceived product quality was significantly positively associated with attitude

toward a product (β1 = .77, t = 46.28, p =.000) and behavioral intentions (β3 =.40, t = 8.09,

p =.000). The results also provided the evidence of negative effects of perceived risk on

attitude toward a product (β2 = -.13, t = -5.00, p =.000) and behavioral intentions (β4 = -

.29, t = -10.20, p =.000). In addition, the results showed that attitude toward a product

significantly and positively influenced behavioral intentions (β5 = .26, t = 5.28, p =.000).

Therefore, H6 through H10 were also supported in the alternative model.

148

The SEM model also revealed indirect and direct effects from perceived quality to

behavioral intentions and from perceived risk to behavioral intentions. The total effect of

perceived product quality on behavioral intentions (β3 + β1β5) is calculated by the sum of the direct effect (β3) and the indirect effect (β1β5), and the total effect of perceived risk on

behavioral intentions (β4 +β2β5) consists of the direct effect (β4) and the indirect effect

(β2β5). Results found that perceived product quality influenced behavioral intentions directly (β3 =.30) and indirectly (β1β5 =.20) through attitude toward a product, and the

effect of perceived risk had a stronger direct effect (β4= -.29) as compared to an indirect

effect on behavioral intentions through attitude toward a product (β2β5 = -.03).

From the SEM model of the study, it is also possible to find indirect effects from

perceived haptic imagery or perceived interactivity to attitude toward a product and

indirect effects from haptic imagery or perceived interactivity to behavioral intentions.

Results revealed an indirect effect for perceived haptic imagery on attitude toward a product (γ1β1 + γ2β2 = .26) through perceived product quality (γ1β1 =.25) and perceived

risk (γ2β2 = .01) and the indirect effect of perceived interactivity on attitude toward a

product (γ3β1 + γ4β2 = .27) through perceived product quality (γ3β1 =.25) and perceived

risk (γ4β2 = .02). Thus, perceived haptic imagery indirectly influenced attitude toward a

product mostly through perceived product quality rather than through perceived risk.

Perceived product quality was also a stronger mediator than perceived risk in terms of

influencing the relationships and between perceived interactivity indirectly and attitude

toward a product. The results found that perceived haptic imagery influenced behavioral

intentions indirectly (γ1β3 + γ1β1 β5 + γ2β2 β5 + γ2β4 = .22) through perceived product

quality, perceived risk and attitude toward a product. The results also revealed that

149

perceived interactivity affected behavioral intentions indirectly (γ4β4 + γ4β2 β5 + γ3β1 β5 +

γ3β3 = .26) through perceived product quality, perceived risk and attitude toward a product (See Table 5.27).

Para. Est. SE t Structural path Haptic imagery (ξ1) → Perceived product quality (η1) γ1 .33 .04 9.23***

Haptic imagery (ξ1) → Perceived risk (η2) γ2 -.05 .05 -1.07 Interactivity (ξ1) → Perceived product quality (η1) γ3 .32 .03 10.14*** Interactivity (ξ2) → Perceived risk (η2) γ4 -.19 .04 -4.73*** Product quality (η1) → Attitude toward a product (η3) β 1 .77 .02 46.28*** Perceived risk (η2) → Attitude toward a product (η3) β 2 -.13 .03 -5.00*** Product quality (η1) → Behavioral Intentions (η4) β 3 .40 .05 8.09*** Perceived risk (η2) → Behavioral Intentions (η4) β 4 -.29 .03 -10.20*** Attitude toward a product (η3) → Behavioral Intentions (η2) β 5 .26 .05 5.28*** Measurement model Haptic imagery (ξ1) → H1 λx1 .68 .03 28.30*** Haptic imagery (ξ1) → H2 λx2 .64 .02 26.41*** Haptic imagery (ξ1) → H3 λx3 .64 .02 26.41*** Interactivity (ξ2) → IN1 λx4 .81 .01 71.77*** Interactivity (ξ2) → IN2 λx5 .93 .01 120.08*** Interactivity (ξ2) → IN3 λx6 .86 .01 87.77*** Product quality (η1) → PQ1 λy1 .66 .02 35.34*** Product quality (η1) → PQ2 λy2 .63 .02 32.11*** Product quality (η1) → PQ3 λy3 .89 .01 71.69*** Perceived risk (η2) → PR2 λy4 .70 .03 27.44*** Perceived risk (η2) → PR3 λy5 .56 .03 21.12*** Perceived risk (η2) → PR4 λy6 .67 .03 26.07*** Attitude toward a product (η3) → A2 λy7 .88 .01 112.54*** Attitude toward a product (η3) → A3 λy8 .89 .01 121.14*** Attitude toward a product (η3) → A4 λy9 .92 .01 141.36*** Behavioral Intention (η4) → I1 λy10 .88 .01 65.86*** Behavioral Intention (η4) → I2 λy11 .59 .02 28.26*** Behavioral Intention (η4) → I3 λy12 .72 .02 42.89*** Note. *p<.05, **p<.01, ***p<.001.

Continued

Table 5.26. Results from the SEM for Testing the Alternative Model

150

Table 5.26. continued

Factor covariance Haptic imagery (ξ1) ↔ Interactivity (ξ2) φ1 .44 .03 14.75*** Error variance/covariance ζ1 ↔ ζ1 Ψ1 .69 .03 25.50*** ζ2 ↔ ζ2 Ψ2 .95 .01 66.24*** ζ3 ↔ ζ3 Ψ3 .37 .02 15.35*** ζ4 ↔ ζ4 Ψ4 .46 .03 17.15*** δ1 ↔ δ1 Θδ1 .54 .03 16.77*** δ2 ↔ δ2 Θδ2 .59 .03 18.81*** δ3 ↔ δ3 Θδ3 .61 .03 20.08*** δ4 ↔ δ4 Θδ4 .34 .02 18.62*** δ5 ↔ δ5 Θδ5 .14 .01 9.41*** δ6 ↔ δ6 Θδ6 .27 .02 15.97*** ε1 ↔ ε1 Θε1 .57 .02 23.36***

ε2 ↔ ε2 Θε2 .61 .03 24.73*** ε1 ↔ ε2 Θε1,2 .24 .02 12.11***

ε3 ↔ ε3 Θε3 .21 .02 9.42***

ε4 ↔ ε4 Θε4 .51 .04 14.44***

ε5 ↔ ε5 Θε5 .69 .03 23.39***

ε6 ↔ ε6 Θε6 .56 .04 16.39*** ε7 ↔ ε7 Θε7 .23 .01 16.44***

ε8 ↔ ε8 Θε8 .21 .01 15.66***

ε9 ↔ ε9 Θε9 .16 .01 13.68***

ε10 ↔ ε10 Θε10 .23 .02 9.56***

ε11 ↔ ε11 Θε11 .65 .03 26.62*** ε12 ↔ ε12 Θε12 .48 .02 19.64*** Model fit Chi-square 496.67 (df = 124, p <.000) RMSEA .048 90 % C.I. (.044; .053) p for the test of perfect fit .000 p for the test of close fit .751 TLI .96 NFI .96 IFI .97

151

Indirect Direct Total Dependent variables Predictor variables effects effects effects Attitude toward a product Perceived haptic imagery .26 .26 Perceived interactivity .27 .27 Perceived quality .77 .77 Perceived risk -.13 -.13

Behavioral intentions Perceived haptic imagery .22 .22 Perceived interactivity .26 .26 Perceived quality .20 .30 .50 Perceived risk -.03 -.29 -.32 Attitude toward a product .26 .26

Table 5.27. Decomposition of Direct, Indirect, and Total Effects for the Alternative

Model.

152

CHAPTER 6

DISCUSSION AND CONCLUSION

The purpose of this chapter is to summarize results of the study and to discuss

implications, limitations and future research. Firstly, this chapter presents a summary of

research findings from the original model testing and the alternative model testing (See

Table 6.1). The second section of this chapter addresses theoretical and managerial

implications. Finally, this chapter discusses limitations of this study and suggests future

research.

Summary and Conclusion

Non-store retailing has dramatically increased and attracts customers by providing opportunities for them to have experiences with diverse shopping environments. Apparel and accessories are one of the very popular product categories for successful non-store retailers. However, non-store apparel retailers confront some obstacles. The biggest obstacle is the lack of direct product experience in online and catalog apparel shopping contexts. Online and catalog shoppers have problems assessing apparel products in the information search, product evaluation, and purchase decision situations since they

153

cannot physically examine apparel products and obtain haptic information by directly

touching garments and fabrics. Thus, the study focused on examining the compensatory roles of pictorial and verbal information for haptic information.

Dual coding theory provided the theoretical framework for the proposed model of

the effect of pictorial and verbal information on consumers’ responses. Stimulus-

Organism-Response (S-O-R) paradigm was also used as the theoretical basis for the proposed model which addressed that pictorial and verbal information and shopping contexts (S), through cognitive internal states (O), influence the shopping outcomes (R).

The proposed model of the study was examined by conducting an experiment using a mock apparel website and a mock apparel catalog. Additionally, this research addressed an alternative model based on the results of the originally proposed model testing.

Multivariate analyses of variance and structural equation modeling were used to test both the originally proposed model and the alternative model.

The purposes of the original model in this study were (1) to investigate the

compensatory effect of pictorial and verbal information for haptic information on customers’ internal states (i.e., perceived product quality and perceived risk); (2) to explore if need for touch moderates the effect of pictorial and verbal information on consumer internal states; (3) to investigate situational differences (catalog vs. Internet shopping contexts) on consumers’ internal states; (4) to examine relationships among consumers’ internal states and shopping outcomes. The alternative model was developed to examine the indirect effect of pictorial and verbal information on customer internal states through perceived haptic imagery and perceived interactivity. The purpose of the alternative model was (1) to test the effect of pictorial and verbal information on

154

perceived haptic imagery; (2) to examine NFT as a moderator of the effects of pictorial

and verbal information on perceived haptic imagery; (3) to investigate the effect of

shopping contexts on perceived interactivity; (4) to examine the effects of perceived

haptic imagery and perceived interactivity on perceived product quality and perceived

risk, finally influencing attitudes toward a product and purchase intentions.

The Effect of Pictorial and Verbal Information

The findings from the original model testing did not support the effect of pictorial and verbal information on perceived product quality and perceived risk, whereas the findings from the alternative model testing found that pictorial information and verbal information significantly influenced perceived haptic imagery, affecting perceptions of product quality. The results revealed that it is important whether or not shoppers really perceive haptic imagery when they are exposed to information that uses haptic imagery- provoking strategies.

The role of pictorial and/or verbal information in terms of evoking imagery has been discussed (Childers & Houston, 1984; Fiore & Yu, 2001; Peck & Childers, 2003;

Unnava & Burnkrant, 1991). The previous research has mainly examined the direct effect of pictorial and/or verbal information on recall, recognition, attitude toward a product or a brand, and purchase intentions. However, there might be an opportunity that certain pictorial information generates perceived imagery, finally influencing positive consumer responses. The alternative model of this current study examined this process: pictorial and verbal information influenced perceived haptic imagery, which affected perceived

product quality. The results are consistent with studies of Burns, Biswas and Babin

155

(1993) and Babin and Burns (1997). These researchers have emphasized mediating roles of visual imagery that is generated by marketing stimuli and is found to influence consumer attitudes and intentions.

This research also contributes by providing knowledge of haptic imagery. Mostly, visual imagery has been discussed in the past (Babin & Burns; 1997; Burns, Biswas &

Babin, 1993; Childers & Houston, 1984; Unnava & Burnkrant, 1991). Recently, Peck and

Childers (2003a; 2003b) have examined the effect of pictorial information as a proxy of haptic information by simply comparing between advertisements with a picture and without a picture. This current study focused on the effect of detailed pictorial information (i.e. a picture swatch) since most apparel catalogs and websites basically provide a picture of a product, such as presenting apparel on a mannequin or on a model.

Thus, the study compared between two pictorial information conditions: one contained a basic dress picture (i.e., presentation on a model) with a picture swatch and the other included a basic dress picture without a picture swatch, and found significant differences in perceived haptic imagery. Pictorial information associated with a picture swatch positively influenced perceptions of vividness and imagery elaboration, but did not influence perceived imagery of haptic properties. On the other hand, when people are exposed to pictorial information without a picture swatch, they tend to perceive less haptic imagery (perceptions of vividness and imagery elaboration). The results provide evidence that a picture swatch facilitates the acquisition of haptic information and is used to compensate for haptic information obtained by directly touching a fabric (Peck &

Childers, 2003a, 2003b). Moreover, the more people perceive haptic imagery due to pictorial information, the more positively they evaluate product quality. The

156

compensatory role of visual stimuli (pictorial information) for haptic information is supported by research on the relationships between haptic imagery and visual imagery

(Klatzky et al., 1991; 1993). When people are exposed to high haptic imagery-provoking information (e.g., a fabric swatch), they engage in a haptic imagery process and are likely to retrieve haptic information about similar products (e.g., textures, softness, and smoothness of a product) from memory, which may eliminate the need for direct examination of a product (Klatzky et al., 1991; 1993). During the haptic imagery process, people are likely to evaluate the quality of the product. High perceived haptic imagery tends to be positively related to perceptions of product quality.

Similarly, this study found an effect for verbal information on perceived haptic imagery. When exposed to fabric hand descriptions, people tended to perceive greater vividness, engaged in greater imagery elaboration, and perceived greater haptic imagery.

However, when exposed to style descriptions, people were less likely to perceive haptic imagery (vividness, imagery elaboration, and imagery of haptic properties). In addition, high haptic imagery generated by verbal information tended to positively influence perceptions of product quality. When exposed to fabric hand descriptions people scored higher on all three dimensions of perceived haptic imagery (i.e., vividness, imagery elaboration, imagery of haptic properties). In particular, imagery of haptic properties which refers to mental imagination of pleasant fabric properties (e.g., smoothness, softness, and silkiness) is most likely to be achieved by direct sensory experience. The results suggest that verbal information containing fabric hand descriptions may play a role as a proxy for touch information when touch is unavailable (Peck & Childers, 2003a;

2003b).

157

Individual Differences

The originally hypothesized model predicted that NFT would moderate the effect of pictorial information on perceived risk and perceived product quality and the effect of verbal information on perceived risk and perceived product quality. The alternative model also postulated that NFT would moderate the effect of pictorial information on perceived haptic imagery and the effect of verbal information on perceived haptic imagery. However, the result revealed that there was no moderating effect for NFT. The results are not congruent with Peck and Childers’ study (2003a). Although Peck and

Childers (2003a) found interaction effects between NFT and pictorial information and between NFT and verbal information on confidence in judgments, there were no interactions between them in this current study.

Further investigations about the comparison between the two groups revealed that the high NFT people tended to have more positive perceptions of product quality and higher perceived risk than low NFT people. The result is somewhat inconsistent with past research. According to Peck and Chiders (2003a), when touch is unavailable, high NFT people tend to be more frustrated with an evaluation task and have less confidence in their judgment than low NFT people. However, this current study found that high NFT people had more positive evaluations about product quality and higher perceived risk (as would be expected) as compared to low NFT people. In addition, the findings of this research revealed that high NFT people had higher perceived haptic imagery scores and higher perceived interactivity scores than low NFT people, suggesting that high NFT people are more sensitive to website stimuli. Haptic imagery-evoking product information presentations and interactive shopping environmental designs might be

158

effective to attract high NFT people and to compensate for the lack of direct sensory exploration of a product. Although high NFT people had high perceived haptic imagery and positive perceived product quality, they still perceived high risk. It might be possible

that perceived haptic imagery served as a proxy for haptic information when high NFT

people evaluated product quality. However, perceived haptic imagery did not play a role

in reducing perceived risk which could suggest that the risk evoked is not related to the

product, but rather to the process of online shopping. However, since two of the

indicators for perceived risk were product related (psychological risk, performance risk),

this explanation seems unlikely.

Situational Differences

The results from the alternative model found that shopping contexts significantly influenced perceived interactivity, affecting perceived product quality, and perceived risk.

On the other hand, the results from the originally hypothesized model did not support the effect of shopping contexts on perceived product quality and perceived risk. The results found the important role of perceived interactivity in influencing perceived product quality and perceived risk. When people browse and shop online, they naturally perceive higher interactivity as compared to when they browse and shop from a catalog. As compared to online shopping environments, catalog shopping environments are static and passive in terms of providing product information and are unable to create interactive shopping environments (Walsh & Godfrey, 2000). However, online shopping environments contain diverse features which increase consumers’ perceived interactivity.

For example, websites used in the experiment of this research included a certain level of

159

navigation interactivity by providing hyperlinks which allowed participants to freely move back and forth through webpages (Schlosser, 2003). Moreover, the results revealed that the more people perceive interactivity due to shopping contexts, the more positively they evaluate product quality and the less risk they perceive.

Further investigations about perceived interactivity found the interaction effect of pictorial information and shopping contexts on perceived interactivity. Interactive media features and advanced technology functions (e.g., close-ups, zoom function, rotation) facilitate consumers’ perceived interactivity (Keng & Lin, 2000; Ko et al., 2005;

Schlosser, 2003). Websites which contain a picture swatch allow consumers to click on a picture swatch icon. While shoppers are clicking on the icon and exploring a picture swatch, they may have higher perceived interactivity and virtual experience as compared to when they are searching a website which does not contain a picture swatch. This study verified that the level of interactive features of websites plays a role in increasing customers’ perceived interactivity. Increasing the interactive and effective features online is important because it has a potential to influence positive evaluations of product quality and reduce perceived risk. On the other hand, since catalogs cannot provide any interactive functions (e.g., closer view and hyperlink), catalogs with a picture swatch did not contribute to increasing perceived interactivity.

The results also found the interaction effect of NFT and shopping contexts on

perceived imagery. High NFT people tended to perceive more vivid imagery when they

browsed online as compared to when they browsed in the catalog, whereas low NFT

people were less likely to perceive vivid imagery in both catalog and online shopping

160

contexts. In order to obtain haptic information, high NFT people were likely to actively engage in information seeking processes by using interactive features of websites.

The Relationships among Consumers’ Internal States and Shopping Outcomes

Structural equation modeling was used to examine the second parts of the original model and the alternative model which addressed the relationships among consumers’ cognitive internal states and shopping outcomes. The original model proposed that positively perceived product quality and less perceived risk have a positive impact on attitude toward a product and purchase intentions. The original model predicted that a favorable attitude toward a product positively influences behavioral intentions. The alternative model added two more variables (i.e., perceived haptic imagery and perceived interactivity) to the original model. The alternative model predicted that highly perceived

haptic imagery generated by pictorial and verbal information positively influences

perceived product quality and reduces perceived risk; proposed positive relationships

between perceived product quality and attitude toward a product, between perceived

product quality and behavioral intentions, and between attitude toward a product and

purchase intentions; and expected inverse relationships between perceived risk and

attitude toward a product and between perceived risk and purchase intentions.

The findings of the alternative model supported that high perceived haptic

imagery elicited by pictorial and verbal information positively influenced perceived

product quality but did not influenced perceived risk. Although there was no research

examining the relationship between perceived haptic imagery and perceived product

quality, the study provided evidence of the connection between perceived haptic imagery

161

and perceived product quality. Perceptions of high haptic imagery are likely to influence

positive perceptions of product quality. The results also revealed that perceived

interactivity generated by shopping contexts influenced perceived product quality and

perceived risk. These findings are supported by prior research examining the effect of

interactivity on customers’ affective responses, attitudes toward a product and purchase

intentions (Fiore & Jin, 2003; Ko et al., 2005; Li et al., 2001). However, the effect of

perceived interactivity on perceived product quality and perceived risk as well as the

effect of perceived haptic imagery on perceived product quality has not been tested in

past research. Therefore, this research provides significant evidence that perceived haptic

imagery and perceived interactivity positively influence perceived product quality and

that perceived interactivity is inversely related to perceived risk.

The findings from the originally hypothesized model and the alternative model

supported a positive effect for perceived product quality on attitudes toward a product

and purchase intentions, a negative effect for perceived risk on attitudes toward a product,

and the positive effect of attitudes toward a product on purchase intentions. The results are congruent with prior research (Athanassopoulos et al., 2001; Dick & Basu, 1994;

Vijayasarathy & Jones, 2000; Zeithamal, Berry, & Parasuraman, 1996). Further investigations about the indirect effect of perceived product quality and perceived risk on

purchase intentions were conducted. The results revealed that attitude toward a product

mediated the relationship between perceived product quality and purchase intentions.

However, perceived risk showed a stronger direct influence on purchase intentions rather

than an indirect effect through attitude toward a product. In addition, the alternative

model revealed the indirect effect of perceived haptic imagery on purchase intentions

162

through perceived product quality, perceived risk and attitudes toward a product, and the indirect effect of perceived interactivity on purchase intentions through perceived product quality, perceived risk and attitudes toward a product.

163

Hypotheses Result Original model H1 Pictorial information Æ (a) perceived product quality Not supported Pictorial information Æ (b) perceived risk Not supported H2 Verbal information Æ (a) perceived product quality Not supported Verbal information Æ (b) perceived risk Not supported H3 Pictorial x verbal information Æ (a) perceived product quality Not supported Pictorial information x verbal information Æ (b) perceived risk Not supported H4 Shopping context Æ (a) perceived product quality Not supported Shopping context Æ (b) perceived risk Not supported H5 Pictorial information x NFT Æ (a) perceived product quality Not supported Pictorial information x NFT Æ (b) perceived risk Not supported Verbal information x NFT Æ (a) perceived product quality Not supported Verbal information x NFT Æ (b) perceived risk Not supported H6 Perceived product quality Æ attitude toward a product Supported H7 Perceived risk Æ attitude toward a product Supported H8 Perceived product quality Æ purchase intention Supported H9 Perceived risk Æ purchase intention Supported H10 Attitude toward a product Æ purchase intention Supported

Alternative model H11 Pictorial information Æ haptic imagery Partially supported H12 Verbal information Æ haptic imagery Supported H13 Pictorial information x NFT Æ haptic imagery Not supported Verbal information x NFT Æ haptic imagery Not supported H14 Shopping context Æ interactivity Supported H15 Haptic imagery Æ (a) perceived product quality Supported Haptic imagery Æ (b) perceived risk Not supported H16 Interactivity Æ (a) perceived product quality Supported Interactivity Æ (b) perceived risk Supported H6 Perceived product quality Æ attitude toward a product Supported H7 Perceived risk Æ attitude toward a product Supported H8 Perceived product quality Æ purchase intention Supported H9 Perceived risk Æ purchase intention Supported H10 Attitude toward a product Æ purchase intention Supported

Table 6.1. Summary of Results of the Hypothesis Testing

164

Implications

Theoretical Implications and Contributions

This research provided theoretical implications for understanding the compensatory effect of pictorial and verbal information for haptic information. One important consideration in non-store apparel shopping environments is the lack of direct

and sensory experience. Although many studies have suggested that substitutes for touch

information should be examined, there has been little research on examining whether any

information compensates for haptic information. This study empirically demonstrates that

pictorial information associated with a picture swatch and verbal information associated

with fabric hand descriptions evoke haptic imagery, and verifies that evoked haptic

imagery finally influences behavioral intentions, through perceived product quality and attitude toward a product. When online and catalog shoppers examine apparel products, without touching real garments, it is possible that they perceive haptic information by exploring pictorial and verbal information. It is important to find effective pictorial and verbal information which have the ability to evoke haptic imagery because highly perceived haptic imagery contributes to positive evaluation of product quality, subsequently influencing favorable attitude toward a product and positive behavioral intentions. This study provided evidence that a picture swatch and fabric hand descriptions played a role in stimulating perceived haptic imagery.

The information process of pictorial and verbal information and the effect of imagery evoking information on consumers’ cognitive responses can be explained by dual coding theory. Although past research applied dual coding theory to examine visual

165

imagery (Childers & Houston, 1984; Paivio, 1986), the theory was never used to explore

the haptic imagery process. Dual coding theory postulates that two types of information

(pictorial information and verbal information) are interconnected and independent when

they are activated (Paivio, 1986), and explains that pictorial information is superior to

verbal information because pictorial information tends to be more easily and rapidly facilitated and is more memorable than verbal information (Childers & Houston, 1984;

Paivio, 1986). Although the present study did not address the superiority effect of

pictorial information on perceived haptic imagery, the study found that both pictorial and

verbal information contributed to evoking haptic imagery.

Past research has examined the effects of pictorial and/or verbal information on

visual imagery and on cognitive evaluations (Babin & Burns; 1997; Burns et al., 1993;

Childers & Houston, 1984; Edell & Staellin, 1983; Rossiter & Percy, 1978; Unnava &

Burnkrant, 1991). Researchers have also suggested the need for studies of cross-modality

and multi-sensory images, such as interactions between visual and auditory imagery

(Unnava, Agarwal, & Haugtvedt, 1996), the effect of visual information on spatial

imagery (Blajenkova, Kozhevnikov, & Motes, 2006), and visual-tactile interactions

(Igarashi et al., 2004).

The current study provides useful knowledge for understanding the effect of

visual information on haptic perceptions. According to research on visual-tactile

relationships (Igarash, Kitagawa, & Ichihara, 2004; Klatzky et al., 1991; 1993),

interactions between vision and touch facilitate perceptions of haptic information

(Klatzky et al., 1991). Bryant and Raz (1975) found that when vision only was available

or both vision and touch were available, vision played a role as a proxy of touch.

166

However, when touch only was available, touch did not compensate for vision. Although this current research did not test the compensatory effect of touch for vision, the results are similar to Bryant and Raz’s study. This study found that visual information (pictorial and verbal information) was used to compensate for touch information.

This study also successfully applied the process of haptic imagery and interactivity to the S-O-R paradigm. S-O-R paradigm explains that environmental stimuli

(S) influence consumers’ cognitive and affective evaluations (O), which intervene between environmental stimuli (S) and consumers’ behavioral responses (R) (Mehrabian

& Russell, 1974). The effect of pictorial and verbal information on perceived haptic imagery and the influence of shopping contexts on perceived interactivity can be explained by the relationships between environmental stimuli and cognitive evaluations

(S-O link). The relationships among perceived haptic imagery, perceived interactivity, perceived product quality, perceived risk, attitude toward a product, and behavioral intentions can be addressed by relationships between cognitive evaluations and behavioral responses (O-R link).

The results from this study also revealed evidence for the ability of interactivity to influence consumers’ responses that has been suggested by research on online shopping environments (Keng & Lin, 2000; Ko et al., 2005; Schlosser, 2003). While online shoppers control and manage contents in online shopping environments (e.g., using zoom function, rotation function, close-ups), they are likely to perceive interactivity, influencing active cognitive and affective evaluations and behavioral responses (Keng &

Lin, 2000; Ko et al., 2005; Li et al., 2003; Schlosser, 2003). In this study, when participants were exposed to the websites containing the high level of interactive features

167

(i.e., closer view functions and hyperlink), perceived interactivity was increased and positively influenced perceived product quality, reduced perceived risk, and subsequently affected favorable attitude toward a product and positive behavioral intentions. Therefore, this study provided support for the nature of interactivity in online shopping environments suggested in past research.

Managerial Implications and Contributions

Non-store retailers have emphasized that it is important to find ways to overcome

the inability to directly touch and experience products. Using haptic imagery-evoking information might be a way to present product information effectively in non-store shopping environments. This study suggests that haptic imagery-evoking information strategies can be used to compensate for haptic information and to lead to consumers’ positive cognitive evaluations of product quality, attitudes toward a product, and purchase intentions. The use of fabric hand descriptions and a fabric swatch can be effective haptic imagery-evoking strategies. A content analysis (pretest 4) which examined information of 36 apparel catalogs and 36 apparel websites revealed that most websites (97%) and catalogs (89%) provided fiber and fabric contents (See Table 3.7).

The content analysis examined the availability of information about fabric hand descriptions: texture (e.g., textured, smooth, silky), weight (e.g., lightweight), temperature (e.g., cool), pressure (e.g., soft, flexible). However, detailed fabric hand information was not generally available on websites and catalogs. The results of the content analysis found that about one third of the websites and catalogs provided information about pressure (e.g., softness) of fabric properties, and revealed that a few

168

catalogs and websites provided information about fabric properties of texture, weight, and temperature (See Table 3.7). In addition, the fabric swatch was available in less than one-fifth of the catalogs and fewer than a third of the websites. Apparel catalogers and online retailers need to take into consideration these haptic imagery-evoking types of information. Further examinations about haptic imagery-evoking information might be needed. In order to present information effectively, websites can provide more dynamic and active images by using advanced interactive features. For example, diverse picture sizes and different sides of a product can be presented by using zoom function, rotation function and alternative views. The advanced features have a potential to stimulate haptic imagery. Adding sensory stimulation to catalog pages may also encourage haptic imagery.

For example, adding real fabric samples may be effective in stimulating direct sensory experience in catalogs (Fiore & Yu, 2001).

The findings of this study provide empirical evidence that perceived interactivity facilitates consumers’ cognitive and behavioral responses. The results also contribute to practical implications by confirming the important role of interactive characteristics in online shopping environments. This study suggests that online retailers need to create customized interactive features so that they can potentially increase values of online shopping environments and manner of product presentations and facilitate customers’ positive shopping experiences. The interactive features may help consumers reduce unnecessary shopping tasks and save time by providing efficient and pleasant shopping environments. Thus, utilizing interactive functions in online shopping may enhance effectiveness of product presentations, store image and competitiveness of online retailers.

However, the appropriate level of interactivity might be questionable. It is possible that

169

complex online shopping environments due to too many interactive features might negatively influence consumers’ shopping processes and consumer responses. Advanced multi-media features can influence download times online. Thus study suggests that online retailers need to explore and test the optimized level of interactivity. Compared to online shopping environments, catalogers have limitations in creating interactive shopping environments. In catalog shopping contexts, visual appeal might be critical.

Catalogers need to consider aesthetic elements, such as color, graphic layout and photographic quality, to create pleasant and effective shopping environments. Using unique aesthetic elements and incorporating fabric hand descriptions and high quality graphics may facilitate active information seeking and the haptic imagery process.

Recently, pure catalogers have established their own websites in order to overcome the lack of interactive features, to effectively present product information, and to take advantage of online features. Some catalogers present the traditional catalog formats on their websites and also provide website formats. Thus, when shoppers visit a website, they can browse products through an online catalog and a website at the same time. Online catalogs use the same visual formats as the traditional catalogs in terms of layout, background, and manner of product presentation. Online catalogs might be attractive to consumers who are accustomed to use traditional catalogs. These online catalogs adopted multi-media functions (e.g., hyperlink, zoom function) generally used in online shopping environments in order to provide opportunities for shoppers to have interactive experiences. The functions also allow consumers to freely move from the online catalog to the website, and vice versa. Multi-channel retailing strategies can be

170

effective in providing extensive information, to save transaction costs, and to increase the number of customers and market coverage (Lohse & Spiller, 1998).

Limitations

Several limitations of the study were identified. Firstly, the study employed a

mock website and a mock catalog and certain aspects of the website and the catalog were

manipulated in order to conduct an experiment. Although the mock website was

developed in a manner to simulate realistic online shopping contexts, the mock website

allowed limited functions (e.g., hyperlink and closer view) and excluded other

technological functions (e.g., search function, email) available on general websites. In

addition, the mock catalog was presented on a computer screen instead of using a more

realistic paper version of the mock catalog. A manipulation check of the reality of

shopping contexts revealed that the mock catalog was considered less realistic than the

mock website. Therefore, it would be worthwhile to conduct future research by using the

more realistic paper catalog pages. The results of the study might not be directly

applicable to real-world settings and should not be generalized beyond the experimental

environments and manipulations.

The product category employed in the study was a dress. Features of haptic

information associated with other apparel product categories (e.g., sweater, blazer) and

the extent which consumers wish to obtain the haptic information might be different

according to apparel categories. In addition, other product categories, such as notebook,

cell phone, or watch, require different features of haptic information. An individual

171

consumer’s need for touch and information seeking behavior might be different

depending on product categories as well.

The sample of the study was selected by University Registrar among female

college students who enrolled at The Ohio State University. The sample characteristics

(i.e., female young college students) should be taken into consideration. The findings of this research cannot be generalized to other populations (e.g., different age and gender groups).

Future Research

Much fruitful research on haptic imagery can be conducted in the future.

Although the study focused on fabric hand descriptions and a picture swatch, other

detailed product presentations (e.g., zoom function, alternative views) or other product

descriptions might be used as a haptic imagery-evoking strategy. In addition, other

environmental cues or information which are not related to haptic information may play a role in compensating for haptic information. For example, familiar brand name (Park &

Stoel, 2005), attractive webpage background (Stevenson, Bruner, & Kumer, 2000), and

promotion (Honea & Dahl, 2005) might be used as proxies to reduce the lack of direct sensory experience and influence positive evaluation about a product or a brand.

Different apparel product categories (e.g., blazer) should be investigated. The different product types are associated with different amounts of haptic information. For example, when consumers browse for a winter coat, they may consider weight

(heaviness/lightness) of the product. For sweaters, softness/roughness might be taken into

172

consideration. It is also possible that other product categories (e.g., cell phone, watch,

and notebook) might require haptic information. According to Peck and Childers (2003a),

certain product types are associated with instrumental haptic information. For example,

when we purchase a notebook computer, we may want to touch and hold the product in order to examine its weight, and the instrumental touch engages in goal-driven information seeking activities. On the other hand, autoletic haptic information seeking is related to hedonic-oriented activities, such as enjoying sensory experience (Holbrook &

Hirschman, 1982). Future research needs to investigate the compensatory effect for different types of haptic information. Furthermore, NFT and consumer cognitive and behavioral responses should be reexamined in relation to different apparel product types and diverse product categories.

The relationships between visual imagery and haptic imagery should be clearly examined. Visual imagery and haptic imagery occur simultaneously (Kerr, 1983). Thus, when people are exposed to pictorial information or verbal information, they are likely to engage in haptic imagery and visual imagery processes at the same time. Research on visual imagery also has emphasized the important role of visual imagery in active cognitive, affective and behavioral responses. In the study, two experimental conditions for verbal information were fabric hand descriptions and style descriptions. Although this study found that fabric hand descriptions significantly influenced higher perceived haptic imagery than style descriptions, it is possible that style descriptions evoke visual imagery.

Therefore, future research needs to examine the independence and interconnectedness between haptic imagery and visual imagery.

173

This current study focused on haptic information indirectly obtained based on

fabric hand descriptions and a picture swatch rather than through direct touch by hands.

In order to examine more accurate compensatory mechanisms for haptic information,

future research needs to compare outcomes when haptic information is obtained by direct

touch and when haptic information is indirectly gained through pictures and verbal fabric

hand descriptions.

The results of this study reinforce the important role that interactive websites can

play when consumers evaluate product quality, consider possible risks related to a

product purchase, and make a purchase decisions. This research focused on the human-

message interactions which occur when people manipulate contents online (e.g., colors,

zoom function, larger view) (Ko et al., 2005). Along with human-message interactions,

human-human interactions (i.e., reciprocal communication among consumers and

retailers) contribute to the interactivity in online shopping environments. For example, by using email, chat rooms, real time feedback, consumers can interact with online retailers or other consumers. Future research may focus on the effect of human-human interactions on consumers’ responses and examine the different effects of human- message interaction and human-human interaction. Future research should continue to explore both the interactive nature of online shopping and the consequences of interactivity in online shopping environments.

174

BIBLIOGRAPHY

Alreck, P. L., & Settle, R. B. (2002). The hurried consumer: Time-saving perceptions of Internet and catalog shopping. Journal of Database Marketing. 10(1), 25-35.

Anderson, J., & Gerbing, D. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bullutin, 103(3), 411-423.

Anderson, J., Gerbing, D., & Hunter, J. E. (1987). On the assessment of unidimensional measurement: Internal and external consistency, and overall consistency criteria. Journal of Marketing Research, 26, 432-437.

Athanassopoulos, A., Gounaris, S., & Stathakopoulos, V. (2001). Behavioral responses to customer satisfaction : An empirical study. European Journal of Marketing, 35(5/6), 687-707.

Baker, J., Parasuraman, A., Grewal, D., & Voss, G. B. (2002). The influence of multiple store environment cues on perceived merchandise value and patronage intentions. Journal of Marketing, 66, 120-141.

Babin, L. A., & Burns, A. C. (1997). Effects of print ad pictures and copy containing instructions to imagine on metal imagery that mediates attitudes. Journal of Advertising, 26(3), 33-44.

Baggozzi, R. P. (1981). Evaluating structural equation models with unobservable variables and measurement error: A comment. Journal of Marketing Research, 18, 375-381.

Bagozzi, R. P., & Phillips, L. W. (1991). Assessing construct validity in organizational research. Administrative Science Quarterly, 36, 421-458.

Baggozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16(1), 74-94.

Baggozzi, R. P., Yi, Y., & Phillips, L. (1982). Representing and testing organizational theories: A holistic construal. Administrative Science Quarterly, 27, 459-489.

175

Biocca, F., Kim, T., & Levy, M. (1995). The vision of visual reality. In F. Biocca & M. R. Levy (Eds.), Communication in the age of virtual reality (pp. 3-14). Hillsdale, NJ: Lawrence Erlbaum.

Blackwell, R. D., Miniard, P. W., & Engel, J. F. (2001). Consumer behaviors (9th ed.). FL: Harcout College Publishers.

Blajenkova, O. Kozhevnikov, M., & Motes, M. A. (2006). Object-spatial imagery: A new self-report imagery questionnaire. Applied Cognitive Psychology, 20, 239-263.

Bloemer, J., Ruyter, K., & Wetzels, M. (1999). Linking perceived service loyalty: A multi-dimensional perspective. European Journal of Marketing, 33(11/12), 1082- 1106.

Bollen, K. A. (1989). Structural equations with latent variables. New York: Wiley.

Brohan, M. (2006). The top 500 guide overview. In K. T. Peters (Ed.), Internet retailer: Strategies for multi-channel retailing 2006 edition top 500 guide (pp. 6-16). NY: Vertical Web Media.

Browne, M., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods and Research, 21(2), 230-258.

Bruner, G. C. (1998). Standardization and justification: Do aad scales measure up? Journal of Current Issues and Research in Advertising, 20(1), 1-18.

Bryant, P., & Raz, I. (1975). Visual and tactual perception of shape by young children. Developmental Psychology, 11, 525-526.

Burns, A. C., Biswas, A., & Babin, L. A. (1993). The operation of visual imagery as a mediator of advertising effects. Journal of Advertising, 22(2), 71-84.

Burke, R. R. (1997). Do you see what I see? The future virtual shopping. Academy of Marketing Sciences, 25(4), 352-360.

Cairns, E., & Coll, P. (1977). The role of visual imagery in visual, tactual, and cross- modal matching. British Journal of Psychology, 68, 213-214.

Cantwell, E. (2004). Forrester research projects US ecommerce to hit nearly $230 billion in 2008. Forrester. Retrieved January 1, 2005, from: http://www.forrester.com/ER/Press/Release/0,1769,823,00.html

Cases, A. (2002). Perceived risk and risk-reduction strategies in Internet shopping. The International Review of Retail, Distribution and Consumer Research, 12(4), 375- 394.

176

Cattell, R. (1966). The scree test for the number of factors. Multivariate Behavioral Research, 1, 245-276.

Childers, T. L., & Houston, M. (1984). Conditions for a picture superiority effect on consumer memory. Journal of Consumer Research, 11, 643-655.

Childers, T. L., Houston, M. J., & Heckler, S. E. (1985). Measurement of individual differences in visual versus verbal information processing. Journal of Consumer Research, 12, 125-458.

Consumer Report Buying Guide (2006). Shopping strategies to get best values. 13-20.

Cox, D. F. (1967). Risk handling in consumer behavior. In D. Cox (Ed.), Risk taking and information handling in consumer Behavior, (pp. 34-81). Boston: Harvard University.

Cunningham, S., (1967). The major dimensions of perceived risk. In D. Cox (Ed.), Risk taking and information handling in consumer behavior, (pp. 84-86). Boston: Harvard University.

Davis, L. I. (1985). Effects of physical quality and brand labeling on perceptions of clothing quality. Perceptual and Motor Skills, 61, 671-677.

Dick, A. S., & Basu, K. (1994). Customer loyalty: Toward an integrated conceptual framework. Journal of the Academy of Marketing Science, 22(1), 99-113.

Dowling, G. R., & Staelin, R. (1994). A model of perceived risk and intended risk- handling activity. Journal of Retailing, 59(1), 119-134.

Drever, J. (1952/1973). A dictionary of psychology. Baltimore, MD: Penguin Books.

Eastlick, M. A., & Feinberg, R. A. (1999). Shopping motives for mail catalog shopping. Journal of Business Research, 45, 281-290.

Eckman, M., Damhorst, M. L., & Kadolph, S. J. (1990). Toward a model of the in-store purchase decision process: Consumer use of criteria for evaluating women’s apparel. Clothing and Textiles Research Journal, 8(2), 13-22.

Edell, J. A. & Staelin, R. (1983). The information processing of pictures in print advertisements. Journal of Consumer Research, 10, 45-61.

Ellen, P. S., & Bone, P. F. (1991). Measuring communication-evoked imagery processing. In R. H. Holman & M. R. Solomon (Eds.), Advances in Consumer Research, (pp. 806-812). Provo, UT: Association for Consumer Research.

177

Ellis, B. C., & Garnsworthy, R. K. (1980). A review of techniques for the assessment of hand. Textiles Research Journal, 4, 231-238.

Eroglu, S. A., Machleit, K. A., & Davis, L. M. (2001). Atmospheric qualities of online retailing: A conceptual model of implications. Journal of Business Research, 54, 177-184.

Eroglu, S. A., Machleit, K. A., & Davis, L. M. (2003). Empirical testing of a model of online store atmospherics and shopper responses. Psychology and Marketing, 20(2), 139-150.

Fetherman, M. S., & Pavlou, P. A. (2003). Predicting e-services adoption: A perceived risk facets perspective. International Journal of Human-Computer Studies, 59, 451- 474.

Fiore, A. M., & Jin, H. (2003). Influence of image interactivity on approach responses toward an online retailer. Internet Research: Electronic Networking Applications and Policy, 13(1), 38-48.

Fiore, A. M., & Yu, H. (2001). Effects of imagery copy and product samples on response toward the product. Journal of Interactive Marketing, 15(2), 36-46.

Fishbein, M. A., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.

Forsythe, S. M. (1991). Effect of private, designer, and national brand names on shoppers’ perception of apparel quality and price. Clothing and Textiles Research Journal, 9(2), 1-6.

Forsythe, S., Presley, A. B., & Caton, K. W. (1996). Dimensions of apparel quality influencing consumers’ perceptions. Perceptual and Motor Skills, 83, 299-305.

Forsythe, S. M., & Shi, B. (2003). Consumer patronage and risk perceptions in Internet shopping. Journal of Business Research, 56, 867-875.

Gallo, R., & McAlister, J. (2003). The top 50 e-retailers. Retailforward.com, August, 1-10.

Garvin, D. A. (1987). Competing on the eight dimensions of quality. Harvard Business Review, 65(6), 101-109.

Gehrt, K. C., & Shim, S. (1998). A shopping orientation segmentation of French consumers: Implications for catalog marketing. Journal of Interactive Marketing, 20(4), 34-46.

178

Gibson, J. J. (1962). Observations on active touch. Psychological Review, 69, 477-491.

Gibson, J. J. (1966). The senses considered as perceptual systems. Boston: Houghton Mifflin.

Grant, L. (2002, October 9). Here come the catalogs. USA Today Retrieved October 10, 2005, from http://www.usatoday.com/money/industries/retail/2002-10-08- holiday-catalogs_x.htm

Grewal, D., Munger, J., Iyer, G., & Levy, M. (2003). The influence of Internet retailing factors on price expectations. Psychology and Marketing, 20(6), 477-493.

Guest, G. (2006, February 21). Shoppers like buying online, study says. The Columbus Dispatch. Retrieved February 21, 2006, from http://www.columbusdispatch.com/business- story.php?story=dispatch/2006/2/21/20060221-C2-02.html

Ha, Y., Kwon, W., & Lennon, S. J. (in press). Online visual merchandising (VMD) of apparel websites. Journal of Fashion Marketing and Management.

Hatch, K., & Roberts, J. (1985). Use of intrinsic and extrinsic cues to assess textile product quality. Journal of Consumer Studies and Home Economics, 9, 341-357.

Hoffman, D. L., & Novak, T. P. (1996). Marketing in hypermedia computer-mediated environments: Conceptual foundations. Journal of Marketing, 60, 50-68.

Hoffman, D. L., Novak, T. P., & Chatterjee, P. (1995). Commercial scenarios for the web: Opportunities and challenges. Journal of Computer-Mediated Communication, 3 (1). Retrieved June 1, 2006, from http://jcmc.indiana.edu.proxy.lib.ohio-state.edu/vol1/issue3/hoffman.html

Holbrook, M. B., & Hirschman, E. C. (1982). The experimental aspects of consumption: Consumer fantasies, feeling, and fun. Journal of Consumer Research, 9, 132-140.

Holbrook, M. B., & Moore, W. L. (1981). Feature interactions in consumer judgments of verbal versus pictorial presentations. Journal of Consumer Research, 8, 103-113.

Honea, H., & Dahl, D. W. (2005). The promotion affect scale: Defining the affective dimensions of promotion. Journal of Business Research, 58, 543-551.

Hong, W., Thong, J. Y. L., & Tam, K. Y. (2004). Designing product listing pages on e- commerce websites: An examination of presentation mode and information format. International Journal of Human-Computer Studies, 61, 481-503.

179

Huizingh, E. K. R. E. (2000). The content and design of web sites: An empirical study. Information and Management, 37, 123-134.

Igarashi, Y., Kitagawa, N., & Ichihara, S. (2004). Vision of a pictorial hand modulates visual-tactile interactions. Cognitive, Affective, & Behavioral Neuroscience, 4(2), 182-192.

Internet Retailer (2004, March 12). Women make up nearly two-thirds of online apparel and beauty shoppers. Retrieved October 10, 2005, from http://internetretailer.com/dailynews.asp?id=11521.

Internet Retailer (2005). Top 400 guide. NY: Vertical Web Media.

Jacoby, J., Olson, J. C., & Haddock, R. A. (1971). Price, brand name and product composition characteristics as determinants of perceived quality. Journal of Applied Psychology, 55, 570-579.

Joines, J. L., Scherer, C. W., & Scheufele, D. A. (2003). Exploring motivations for consumer web use and their implications for e-commerce. Journal of Consumer Marketing, 20(2), 90-108.

Jones, J. M., & Vijayasarathy, L. R. (1998). Internet consumer catalog shopping: Findings from an exploratory study and directions for future research. Internet Research: Electronic Networking Applications and Policy, 8(4), 322-330.

Kaiser, H. (1960). The application of electronic computers to factor analysis. Educational and Psychological Measurement, 20, 141-151.

Katz, D. (1989). The world of touch. Hillsdale, NY: Erlbaum.

Kaufman-Scarborough, C., & Lindquist, J. D. (2002). E-shopping in a multiple channel environment. Journal of Consumer Marketing, 19(4), 333-350.

Keng, C., & Lin, H. (2006). Impact of telepresence levels of Internet advertising effects. Cyber Psychology & Behavior, 9(1), 82-94.

Kerr, N. H. (1983). The role of vision in visual imagery experiments: Evidence from the congenitally blind. Journal of Experimental Psychology: General, 112, 265-277.

Kim, M. J., Kim, J-H., & Lennon, S. J. (in press, 2006). Online service available on apparel retail websites: An E-S-QUAL approach. Managing Service Quality, 16(1).

180

Kim, M, & Lennon, S. J. (2000). Television shopping for apparel in the United States: Effects of perceived amount of information on perceived risks and purchase intentions. Family and Consumer Sciences Research Journal, 28(3), 301-330.

Kisielius, J., & Sternthal, B. (1984). Detecting and explaining vividness effects in attitudinal judgment. Journal of Marketing Research, 21, 54-64.

Klatzky, R. L., Lederman, S. J., & Matula, D. E. (1991). Imagined haptic exploration in judgments of object properties. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17(2), 314-322.

Klatzky, R. L., Lederman, S. J., & Matula, D. E. (1993). Haptic exploration in the presence of vision. Journal of Experimental Psychology: Hyman Perception and Performance, 19(4), 726-743.

Ko, H., Cho, C., & Roberts, M. S. (2005). Internet uses and gratifications: A structural equation model of interactive advertising. Journal of Advertising, 34(2), 57-70.

Koering, S. K. (2003). E-scapes: The electronic physical environment and service tangibility. Psychology and Marketing, 20(2), 151-167.

Korgaonkar, P. K. (1984). Consumer shopping orientation, non-store retailers and consumers’ patronage intentions: A multivariate investigation. Journal of the Academy of Marketing Science, 12, 11-22.

Kosslyn, S. M. (1980). Image and mind. Cambridge, MA: Harvard University Press.

Kwon, W. (2005). A model of reciprocal effects of multi-channel retailers' offline and online brand images: Application to multi-channel specialty apparel retailing. Unpublished doctoral dissertation, The Ohio State University, Columbus.

Kwon, Y. H., Paek, S. L., & Arzeni, M. (1991). Catalog vs. non-catalog shoppers of apparel: Perceived risks, shopping orientations, demographics, and motivations. Clothing and Textiles Research Journal, 10(1), 13-19.

Lea-Greenwood, G. (1998). Visual merchandising: A neglected area in UK fashion marketing? International Journal of Retail and Distribution Management, 26(8), 324-329.

Li, H., Daugherty, T., & Biocca, F. (2001). Characteristics of virtual experience in electronic commerce: A protocol analysis. Journal of Interactive Marketing, 15(3), 13-30.

181

Li, H., Daugherty, T., & Biocca, F. (2002). Impact of 3-D advertising on product knowledge, brand attitude, and purchase intention: The mediating role of Presence. Journal of Advertising, 3, 43-57.

Li, H., Daugherty, T., & Biocca, F. (2003). The role of virtual experience in consumer learning. Journal of Consumer Psychology, 13(4), 395-407.

Li, Y. N., Tan, K. C., & Xie, M. (2002). Measuring web-based service quality. Total Quality Management, 13(5), 685-700.

Lohse, G. L., & Spiller, P. (1998). Electronic shopping. Communication of the ACM, 41(7), 81-7.

Lu, Y., & Rucker, M. (2006). Apparel acquisition via single vs. multiple channels: College students’ perspectives in the US and China. Journal of Retailing and Consumer Services, 13, 35-50.

Lusch, R., & Brown, J. (1996). Interdependency, contracting, and relational behavior in marketing channels. Journal of Marketing, 60(4), 19-38.

Lutz, K. A., & Lutz, R. J. (1977). The effects of interactive imagery and learning: Application to advertising. Journal of Applied Psychology, 62, 493-498.

Lutz, K. A., & Lutz, R. J. (1978). Imagery-eliciting strategies: Review and implications of research. Advances in Consumer Research, 5(1), 611-620.

MacInnis, D. G., & Price, L. L. (1987). The role of imagery in information processing: Review and extensions. Journal of Consumer Research, 13, 473-491.

MacKinnon, D. P., Warsi, G., & Dwyer, J. H. (1995). A simulation study of mediated effect measures. Multivariate Behavioral Research, 30(1), 41-62.

Maignan, I., & Lukas, B. A. (1997). The nature and social uses of the Internet: A qualitative investigation. Journal of Consumer Affairs, 31(2), 346-371.

Mathwick, C., Malhotra, N., & Rigdon, E. (2001). Experimental value: Conceptualization, measurement and application in the catalog and Internet shopping environment. Journal of Retailing, 77, 39-56.

Mazzocco, P. (2005). Moderators of the effects of mental imagery on persuasion: The cognitive resources model and the imagery correction model. Unpublished doctoral dissertation, The Ohio State University, Columbus.

McCorkle, D. E. (1990). The role of perceived risk in mail order catalog shopping. Journal of Direct Marketing, 4(4), 26-35.

182

McCabe, D. B., & Nowlis, S. M. (2003). The effect of examining actual products or product descriptions on consumer preference. Journal of Consumer Psychology, 13(4), 431-439.

McGoldrick, P. J. (1990). Retail Marketing. Maidenhead: McGraw-Hill.

McGoldrick, P., & Pieros, C. P. (1998). Atmospherics, pleasure and arousal: The influence of response moderators. Journal of Marketing Management, 14, 173- 197.

McKinney, V., Yoon, K., & Zahedi, F. M. (2002). The measurement of web-customer satisfaction: An expectation and disconfirmation approach. Information Systems Research, 13(3), 296-315.

Mehrabian, A., & Russell, J. A. (1974). An approach to environmental psychology. Cambrige, MA: MIT Press.

Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitudes? Journal of Marketing Research, 18, 318- 332.

Mooy, S. C., & Robben, H. S. J. (2002). Managing consumers’ product evaluation through direct product experience. Journal of Product and Brand Management, 11(7), 432-446.

Olson, J. C., & Jacoby, J. (1972). Cue utilization in the quality perception process. In M. Venkatesan (Ed.), Advances in Consumer Research, Ann Arbor, MI: Association for Consumer Research.

Paivio, A. (1971). Imagery and verbal processes. NY: Holt, Rinehart, & Winston.

Paivio, A. (1975). Perceptual comparisons through the mind’s eye. Memory and Cognition, 3(6), 635-647.

Paivio, A. (1986). Mental representations: A dual coding approach. NY: Oxford University Press.

Paivio, A., & Desrochers, A. (1980). A dual-coding approach to bilingual memory. Canadian Journal of Psychology, 34(4), 388-399.

Park, J. H. (2002). The effect of product presentation on mood, perceived risk, and apparel purchase intention in Internet apparel shopping. Unpublished doctoral dissertation, The Ohio State University, Columbus.

183

Park, J. H., & Stoel. L. (2002). Apparel shopping on the Internet: Information availability on US apparel merchant Web sites. Journal of Fashion Marketing and Management, 6(2), 158-176.

Park, J., & Stoel, L. (2005). Effect of brand familiarity, experience and information on online apparel purchase. International Journal of Retail and Distribution Management, 33(2), 148-160.

Park, J., Lennon, S. J., & Stoel, L. (2005). On-line product presentation: Effects on mood, perceived risk, and purchase intention. Psychology and Marketing, 22(9), 695-719.

Pascale, M. (2000, June 1). 4th annual electronic marketing survey: Online catalogers. Catalog Age. Retrieved July 3, 2006, from http://bg.catalogagemag.com/ar/marketing_th_annual_electronic

Peck, J., & Childers, T. L. (2003a). To have and to hold: The influence of haptic information on product judgments. Journal of Marketing, 67, 35-48.

Peck, J., & Childers, T. L. (2003b). Individual differences in haptic information processing: The “need for touch” scale. Journal of Consumer Research. 30, 430- 442.

Percy, J. R., & Rossiter, J. R. (1983). Effects of picture size and color on brand attitude responses in print advertisement. Advances in Consumer Research, 10, 17-20.

Peterson, R. A., & Merino, M. C. (2003). Consumer information search behavior and the Internet. Psychology and Marketing, 20(2), 99-121.

Ranganatha, C., & Ganapathy, S. (2002). Key dimensions of business-to-consumer web sites. Information and Management, 39, 457-465.

Rao, A. R., & Monroe, K. B. (1988). The impact of prior knowledge on the price- perceived quality relationship. Journal of Consumer Research, 15, 253-264.

Reber, A. S. (1995). Dictionary of psychology. 2nd ed. New York: Penguin Books.

Reichheld, F., & Sasser, W. E. Jr. (1990). Defining quality: alternatives and implications. Academy of Management Review, 13, 419-445.

Roeckelein, J. E. (2004). Imagery in psychology: A reference guide. Westport, CT: Praeger Publishers.

Roselius, T. (1971). Consumer rankings of risk reduction methods. Journal of Marketing, 35, 56-61.

184

Rossiter, J. J. R., & Percy, L. (1978). Visual imaging ability as a mediator of advertising response. In H. K. Hunt, & A. Arbor (Eds.), Advances in Consumer Research (pp. 621-62). MI: Association for Consumer Research.

Rossiter, J. R. (1982). Visual imagery: Applications to advertisint. In A. A. Mitchell, & A. Arbor (Eds.), Advances in Consumer Research (pp. 101-106). MI: Association for Consumer Research.

Schlosser, A. E. (1998). Applying the functional theory of attitudes to understanding the influence of store atmosphere on store inferences. Journal of Consumer Psychology, 7(4), 345-369.

Schlosser, A. E. (2003). Experiencing products in the virtual world: The role of goal and imagery in influencing attitudes versus purchase intentions. Journal of Consumer Research, 30, 184-198.

Schultz, R. (2004, March 1). Multichannel bliss. Direct.com. Retrieved February 21, 2006, from http://directmag.com/mag/marketing_multichannel_bliss/index.html

Seiders, K., Berry, L. L., & Gresham, L. G. (2000). Attention, retailers! How convenient is your convenience strategy? Slogan Management Review, 41(3), 79-89.

Sen, S., Block, L. G., & Chandran, S. (2002). Window displays and consumer shopping decisions. Journal of Retailing and Consumer Services, 9, 277-290.

Sherman, H. (1996). Catalog management. Journal of Direct Marketing, 10(2), 70.

Sherry, C. (2002). Consumer shopping survey. Catalog Age, 18(9), 57 – 59.

Shim, S., Eastlick, M. A., Lotz. S. L., & Warrington, P. (2001). An online prepurchase intentions model: The role of intention to search. Journal of Retailing, 77, 397- 416.

Shim, J. P., Shin, Y. B., & Nottingham. L. (2002). Retailer web site influence on customer shopping: An exploratory study on key factors of customer satisfaction. Journal of the Association for Information Systems, 3, 53-76.

Simpson, L., & Lakner, H. B. (1993). Perceived risk and mail order shopping for apparel. Journal of Consumer Studies and Home Economics, 17, 377-389.

Soufflet, I., Calonnier, M., & Dacremont, C. (2004). A comparison between industrial experts’ and notices’ haptic perceptual organization: A tool to identify descriptors of the handle of fabrics. Food Quality and Preference, 15, 689-699.

185

Spies, K., Hesse, F., & Loesch, K. (1997). Store atmosphere, mood and purchasing behavior. International Journal of Research in Marketing, 14, 1-17.

Srinivasan, S. S., Anderson, R., & Ponnavolu, K. (2002). Customer loyalty in e- commerce: An exploration of its antecedents and consequences. Journal of Retailing, 78, 41-50.

Stell, R., & Paden. N. (1999). Vicarious exploration and catalog shopping: A preliminary investigation. Journal of Consumer Marketing, 16(4), 332-344.

Stevenson, J. S., Bruner II, G. C., & Kumar, A. (2000). Webpage background and viewer attitudes. Journal of Advertising Research, 1-4, 29-34.

Steuer, J. (1992). Defining virtual reality: Dimensions determining telepresence. Journal of Communication, 42(4), 73-93.

Stone-Romero, E. F., Stone, D. L., & Grewal, D. (1997). Development of a multidimensional measure of perceived product quality, Journal of Quality Management, 2(1), 87-111.

Teo, T. S. H., & Yeong, Y. D. (2003). Assessing the consumer decision process in the digital marketplace. Omega, 31, 349-363.

Timmermans, H., & Morganosky, M. (1999). Special issue on direct marketing: Where the old meets the new. Journal of Business Research, 45, 247-248.

Transword Business (2001, August 6). Southern California consumers optimistic about economy. Retrieved June 28, 2006 from http://www.twsbiz.com/twbiz/industrynews/article/0,21214,706547,00.html

Turley, L. W., & Milliman, R. E. (2000). Atmospheric effects on shopping behavior: A review of the experimental evidence. Journal of Business Research, 49, 193-211.

Unnava, H. R., & Burnkrant, R. E. (1991). An imagery-processing view of the role of pictures in print advertisements. Journal of Marketing Research, 33, 226-231.

Unnava, H. R., Agarwal, S. , & Haugtvedt, C. P. (1996). Interactive effects of presentation modality and message-generated imagery on recall of advertising information. Journal of Consumer Research, 23, 81-88.

US Census Bureau News (2005, May 20). Quarterly retail e-commerce sales: 1st quarter 2005. Retrieved June 28, 2006, from http://www.census.gov/mrts/www/data/html/05q1.html

186

Valenzi, E. R., & Andrews, I. R. (1971). Effect of price information on product quality ratings. Journal of Applied Psychology, 55, 87-91.

Vijayasarathy, L. R., & Jones, J. M. (2000). Print and internet catalog shopping: assessing attitudes and intentions. Internet Research: Networking Applications and Policy, 10(3), 191-202.

Walsh, J., & Godfrey, S. (2000). The Internet: A new era in customer service. European Management Journal, 18(1), 85-92.

Ward, M. R., & Lee, M. J. (2000). Internet shopping, consumer search and product branding. Journal of Product & Brand Management, 9(1), 6-20.

Wheatley, J. J., Walton, R. G., & Chiu, J. S. Y. (1977). The influence of prior product experience, price and brand on quality perception. In W. D. Perreault, Jr. (Ed.). Advances in Consumer Research, Ann Arbor, MI: Association for Consume Research.

White, P. D., & Cundiff, E. F. (1978). Assessing the quality of industrial products. Journal of Marketing, 42, 80-86.

Yoh, E., Damhorst, M. L., Sapp, S., & Laczniak, R. (2003). Consumer adoption of the Internet: The case of apparel shopping. Psychology and Marketing, 20(12), 1095- 1118.

Yuille, J. C., & Catchpole, M. J. (1977). The role of imagery in models of cognition. Journal of Mental Imagery, 1, 171-180.

Zeithamal, V. A., Berry, L. L., & Parasuraman, A. (1996). The behavioral consequences of service quality. Journal of Marketing, 60(4), 31-46.

Zeng, X., Koehl, L., Sanoun, M., Bueno, M. A., & Renner, M. (2004). Integration of human knowledge and measured data for optimization of fabric hand. International Journal of General Systems, 33 (2-3), 243-258.

187

APPENDIX A

INVITATION MAIL AND CONSENT FORM

188

Invitation Mail

Dear Participant:

Hello. My name is Minjung Park. I am a doctoral student of Textiles and Clothing in the department of Consumer Sciences at the Ohio State University. You are being asked to participate in this study. Your participation is completely voluntary. The purpose of this research is to investigate online shopper characteristics and behaviors.

After you read the instructions, please feel free to ask any questions you may have. We appreciate your cooperation. This study should take approximately 15 - 20 minutes.

This study is concerned with aggregate data and not with your individual responses. Thus your responses will remain confidential. Your name will not be associated with the data collected. If you do not wish to answer questions, you can refuse to answer them or you can refuse to participate without penalty or repercussion.

Your cooperation and participation in this study are appreciated. If you have any questions or are interested in the results of the study, contact Minjung Park or Dr. Sharron Lennon, listed below.

Sincerely,

Minjung Park Dr. Sharron Lennon, Ph. D. Graduate Student Professor Consumer Sciences Consumer Sciences 614) 688-4234 [email protected] [email protected]

189

Consent Form

Protocol Number: 2006E0302 Investigation: Dr. Sharron Lennon and Minjung Park

I consent to participating in the research.

Dr. Sharron J. Lennon or Minjung Park has explained the purpose of the study, the procedures to be followed, and the expected duration of your participation. Possible benefits of the study have been described, as have alternative procedures, if such procedures are applicable and available.

I acknowledge that I have had the opportunity to obtain additional information regarding the study and that any questions I have raised have been answered to my full satisfaction. Furthermore, I understand that I am free to withdraw consent at any time and to discontinue participation in the study without prejudice to me.

Finally, I acknowledge that I have read and fully understand the consent form. I sign it freely and voluntarily.

Signed: Minjung Park (Principal Investigator or his/her Signed: (Participant) (type your name here) authorized representative)

Please Click "Submit", to continue your survey

Submit

190

APPENDIX B

PRETEST 1

191

Please indicate the importance of touch and fit of the clothing in the decision to purchase the following products. Please, check (√) the number of your response.

Blazers & fitted Touch is Extremely Touch is Not 1 Neutral Jackets Important Important at all 7 6 5 4 3 2 1 Fit is Extremely Fit is Not