• No results found

experiences website

N/A
N/A
Protected

Academic year: 2021

Share "experiences website"

Copied!
39
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

enhancing return visits with an

enjoyable and attractive website

Vera Buring

1384082

December 2008

Museumstraat 7a, 9711 HS Groningen veraburing@gmail.com

06-12 45 63 55

(2)

PREFACE

Sometimes, I found myself lost on the World Wide Web. I became a victim of a pitfall called ‘wilfing’ – an

abbreviation of ‘what was I looking for?’ – but I found that I also enjoyed being sucked into a vortex of information. The citation in the introduction, although fictitious, could be about me. This led me to the concept of flow, as a practical and academical issue regarding the Web. But I also found that it was hard to maintain a state of flow, especially when writing this thesis. Therefore, there are some people I would like to express my thanks to. First, I would to thank my supervisor at the University of Groningen, Jan Kratzer, for his insightful feedback and directions that send me in the right way. Another thanks goes to Wim Biemans, my second assessor.

I would also like to thank my colleagues at NDC mediagroep, for their support and coffee breaks. A special thanks goes to Arjan Breedveld, my supervisor, Linda Bruinsma, my colleague who was wonderful to talk to, and Robert Oosterbaan, for his help distributing the questionnaire. During my internship at a multimedial company, I have had a great time being in the middle of journalism and design on one hand and business on the other hand.

Finally, I would like to thank my family and friends for their support during these last months of my education years, and for getting me in the flow of writing this thesis.

(3)

ABSTRACT

This thesis regards the concept of a flow experience on the World Wide Web, and especially on a particular website. Flow refers to “the state in which people are so involved in an activity that nothing else seem to matter; the

experience itself is so enjoyable that people will do it even at great cost, for the sheer sake of doing it” (Csikszentmihalyi, 1991: 4), and is characterized by a narrowing of the focus of awareness, so that irrelevant perceptions and thoughts are filtered out; loss of self-consciousness; responsiveness to clear goals and unambiguous feedback; and a sense of control over the environment (Csikszentmihalyi, 2000).

In this thesis, the concept of flow follows the work of Webster et al. (1993). Flow is considered by four factors, namely the extent to which (1) the user’s curiosity is aroused during the interaction with a website, (2) the user finds the interaction with the website intrinsically interesting, (3) the user perceives that his or her attention is focused on the interaction with the website, and (4) the user perceives a sense of control over the interaction with the website.

The occurrence of a flow experience also relates to the perceived attractiveness of a website. As flow is an enjoyable feeling a visitor gets when s/he is browsing a website, literature suggests it has positive consequences such as extended duration of visits and repeat visits. It is believed in literature that a flow experience can be prompted by taking into account certain aspects or attributes of a website (Chen et al., 1999; Huang, 2003).

The central constructs in this thesis are a state of flow experienced on a website, the attractiveness of a website, and the intention to return to a website. Several content and design attributes are regarded for creating a website that is attractive and enjoyable. A website attribute is an aspect of a website, where content attributes refer to the kind of information presented on the website and design attributes to the way the information is presented on the website, or the layout of the website (Huizingh, 2000).

In this thesis, the following research question is posed: which website content and design attributes contribute to an attractive website where a flow experience can occur, so that visitors will return to the website. Central in this thesis is the repositioning of the website of a multimedial living related brand. The purpose of this thesis is the website’s redevelopment, regarding the content and design of the website.

Considered are eight content attributes divided in three groups, namely the presence of regional content on a website, the source of the content (which could originate from a website’s editors, third parties or the website’s users) and four interactive features (the presence of a Web poll, a forum, a Web shop and the option to comment on an article). Also considered are seven design attributes divided in three groups, namely the level of personalization on a website, the presentation style (whether images and videos are used instead of traditional text), and the navigation type (the way content pages are interlinked).

These website attributes are hypothesized to positive relate to the occurrence of a flow experience on a website and the perceived attractiveness of a website. Also hypothesized are the impact of a website’s attractiveness on flow, and the relation of attractiveness and flow on the intention to return to that website.

(4)

As a lot of websites were only mentioned a couple of times, these seventeen most mentioned websites, form the basis for study 2. These websites are individually analyzed by the researcher, whether each of the fifteen content and design attributes is present on a unique website.

It is analyzed whether the presence of the content and design attributes impacts the occurrence of a flow experience and the perceived attractiveness of a website. These first four hypotheses are confirmed. Most of these relations are significant, but several show to influence attractiveness and/or the flow experience in a negative way, instead of the hypothesized positive relation.

Among the interesting results are that content generated by users is found to decrease both a website’s attractiveness and the occurrence of a flow experience. Also, the presence of information customized to a visitor shows to enhance both flow and attractiveness. Furthermore, the use of videos and images as a presentation style is not related to attractiveness, as was suggested in literature, and shows to decrease flow.

The central hypotheses are all confirmed, as suggested in literature. It shows that the perceived attractiveness of a website enhances the occurrence of a flow experience, and that the intention to return to a website is caused by both attractiveness and flow.

(5)

CONTENTS

1. INTRODUCTION ...6 2. THEORETICAL FRAMEWORK ...8 2.1 Website Attributes... 8 Content Attributes... 8 Design Attributes ... 9 2.2 Attractiveness...11 2.3 Flow Construct ...11

Flow as a Mental Mode...11

Representations of the Flow Construct ...12

2.4 Flow Factors ...13 2.5 Intention to Return ...15 3. METHODOLOGY ... 16 3.1 Case Description ...16 3.2 Study Design ...16 3.3 Data Collection...16 3.4 Measures...17 Content Attributes...18 Design Attributes ...18

Attractiveness, Flow Experience, and Intention to Return...19

4. ANALYSIS... 22

4.1 Descriptive Statistics...22

4.2 Results ...23

Content Attributes...23

Design Attributes ...23

Attractiveness, Flow Experience, and Intention to Return...25

5. DISCUSSION AND CONCLUSION... 26

5.1 Impact of Content Attributes ...26

5.2 Impact of Design Attributes...28

5.3 Enhancing the Intention to Return...29

5.4 Recommendations ...30

5.5 Limitations ...31

5.6 Suggestions for Further Research ...31

REFERENCES...32

(6)

1. INTRODUCTION

“Great websites are not about navigating content, but staging experience” Mihaly Csikszentmihalyi on online flow experiences (Geirland, 1996)

Imagine this setting: It is 9:45 p.m. A young female starts up her laptop, checking her mail account. A message from a friend abroad, who writes about a monument he recently saw. When she looks at the photos he sent along, she gets interested in what happened there. She searches the Web for some background information. She opens the first hit on Google. On the website, some information is cited about the occurrence, but she is curious for more information. She goes back to the search results and opens another website. She finds this website more attractive. She reads, navigates the website and in her head the whole story connects like a puzzle. Then she looks at the clock. It reads 12:10 a.m. She goes to bed, remembering to revisit the website the next day.

One might want to know what made the last website so enjoyable that she stayed for such a long time. In this thesis, it is hypothesized that she experienced a state of flow. Originating from the field of psychology, flow refers to “the state in which people are so involved in an activity that nothing else seem to matter; the experience itself is so enjoyable that people will do it even at great cost, for the sheer sake of doing it” (Csikszentmihalyi, 1991: 4).

Flow is a mental state of the individual visitor, which is characterized by a narrowing of the focus of awareness, so that irrelevant perceptions and thoughts are filtered out; loss of self-consciousness; responsiveness to clear goals and unambiguous feedback; and a sense of control over the environment (Csikszentmihalyi, 2000). Csikszentmihalyi (1991, 2000) studied this optimal experience known as flow in physical activities like sports and playing chess. However, the concept of flow was later adopted to research in computer environments – for example the works of Hoffman and Novak (1996) and Novak et al. (2000). In Chen et al.’s (1999: 597) research on flow experiences in an online environment, respondents reported feelings of “inspired involvement”, “lost self-consciousness”, “exciting and fascinated”, and “timelessness” in a flow state. These descriptions are similar to those experienced in other research.

Flow however is not a constant state; people move in and out of flow. Flow is viewed as both a trait and a state (Hoffman and Novak, 1996). As traits are “comparatively stable characteristics of individuals that are relatively invariant to situational stimuli”, states are “affective or cognitive episodes that are experienced in the short run and fluctuate over time”, and, unlike traits, “can be influenced by situational factors and the interaction between the person and the situation” (Webster and Martocchio, 1992: 203). A flow experience is autotelic, where the experience is an end in itself (Csikszentmihalyi, 2000; Hoffman and Novak, 1996). Therefore, a person who has an autotelic trait is more likely to experience a state of flow for a given activity.

Flow as an online customer experience is important for critical marketing tasks, such as extended duration of visits and repeat visits (Novak et al., 2000). When one is in a state of flow, s/he wants to extend this experience and repeat it (Csikszentmihalyi, 2000). This is also confirmed by Chen et al. (2000), by stating that one wants to duplicate the enjoyable flow experience as much as possible. In recent literature, the intention to return to the website is often researched as a consequence of a flow experience (e.g. Koufaris, 2002; Siekpe, 2005; Skadberg and Kimmel, 2004). Therefore, this thesis proposes that the occurrence of flow on a website causes one to visits to this website again.

Central in this thesis is the case of a multimedial living related brand, consisting of a fair, a supplement in two newspapers, and a website. A glossy magazine, introduced together with the website, was cancelled after a few editions, due to economic reasons. This thesis focuses on the brand’s website, which has to serve as an online

(7)

Shortly after the start of this research, a brand manager was appointed to oversee the brand. The brand’s website used to focus on general-economic news facts and extensions to articles in the newspaper supplement. Due to the repositioning of the brand, the focus will be more on living related inspirational topics. The time one spends on such a website, will rely on the feeling one gets when s/he gets inspired by the website’s content. Therefore, in this context, the occurrence of a flow experience is a relevant topic.

The purpose of this thesis is to recommend the brand’s management on the redevelopment of the website, to attract more visitors to the website. Not included in this thesis is the organizational implementation, which relates to the question which editors are responsible for adding content to the website. Also excluded are marketing related questions regarding the strategic positioning of the brand.

This thesis will examine which website content and design attributes contribute to an attractive website where a flow experience can occur, so that visitors will return to the website. The following research questions will be answered. First, does the state of flow experienced on a website and the perceived attractiveness of a website make the visitor to return to this website again. Second, to what degree do attributes related to the content and the design of a website create a flow experience. And third, to what degree do content and design attributes of a website enhance the perceived attractiveness.

This thesis is distinctive as it regards the occurrence of a flow experience on a specific website. Flow on the Web has been popular in literature for more than a decade. However, how one website can create a flow experience, is more or less an unexplored topic of interest. In their research, Skadberg and Kimmel (2004) found that when one perceives a website as attractive, the occurrence of a flow experience on that website is more likely. They however did not attempt to research which website attributes contribute to the perception of the attractiveness of the website.

This thesis could serve as a guideline for developing new websites that are enjoyed by its visitors and where they want to return to. As behavioral intentions (in this case, the intention to return to the website) are a consequence of flow, the proposed model can lead to practical applications for businesses as to which website attributes can enable a flow experience.

(8)

2. THEORETICAL FRAMEWORK

2.1 Website Attributes

Central in this thesis is a website, which is an interlinked collection of Web pages with different types of content (e.g. text, images, video), which can be accessed with a Web browser (e.g. Microsoft Internet Explorer). Different pages and websites are connected with hyperlinks. The World Wide Web (hereafter, ‘Web’) refers to all websites and is a way of accessing information over the Internet. It has a tighter definition than the Internet, as the Internet is a global network of computers. However, ‘surfing the Internet’ is general slang for browsing websites.

A website is viewed as a bundle of attributes, which are aspects of a website, which “users see as a bundle of attributes with varying capacities to satisfy their needs” (Huang, 2003: 426). These attributes are separated into content and design attributes, following Huang et al. (2006), Huizingh (2000), and Skadberg and Kimmel (2004). Content refers to the kind of information presented on the website. Design refers to the way the information is presented on the website, or the layout of the website (Huizingh, 2000).

It is assumed that an attractive website can deliberately be developed by taking into account certain website attributes, to create a flow experience so that visitors will revisit the website. It has been researched that visiting websites is such an activity where a flow experience can occur. In Chen et al.’s (1999) research, where the Web is viewed as a multi-activity medium, surfing and locating the desired information was the most named cause of flow. About 60% of the mentioned activities where a state of flow was experienced were browsing the Web looking for information. Also, according to Huang (2003: 436), “a website can be designed with attributes that shape […] flow”.

In the next section, it is attempted to identify which content and design attributes enhance a website’s attractiveness and influence the occurrence of a flow experience. Thereafter, the central theoretical model is described, consisting of the relationship between a flow experience, attractiveness, and the intention to return to a website.

Content Attributes

This thesis distinguishes eight attributes that refer to the content of the website, which are closely related to the positioning of the brand. These attributes are divided into three groups, namely regional scope of the content, uniqueness of the content and interactive features, and are distinguished in different attributes, which are partly suggested by the brand manager and also have its basis in literature.

Regional Content. Since the brand’s unique selling point focuses on the three Northern provinces, the interests of

the target group may be largely geographically related. This item is added because the publisher and the brand focus on regional news. This attribute refers to the focus of the content, to the extent in which the content is related to a geographical area (e.g. the three northern provinces), depending on the type of content.

Uniqueness. The uniqueness of the content is suggested by Koufaris (et al., 2001; Koufaris, 2002) as a source of

(9)

Interactive Features. An interactive website is said to have a positive relationship on a flow experience (Skadberg and Kimmel, 2004). Interactivity refers to the communication process between the website and its users (Huang, 2003). These interactive features refer to certain features that enable the user to exchange information. There are four interactive features differentiated in this thesis, namely a web poll, a forum, a web shop and the option of commenting on an article.

A Web poll is a tool where users’ opinions are collected and graphical reproduced so that the distribution of the opinions is visible. This feature is currently available on websites related to the brand, and could also be implemented on the brand’s website. A forum is an online discussion group, where users can discuss topics by posting messages in a thread. This feature is added to the questionnaire because the brand manager wishes to establish the requirement for a forum on the brand’s website. Currently, a forum is available on the brand’s website, but with no activity from users. A Web shop is an online tool where products can be put in a shopping cart and be purchased. This feature is added to the measurement because the publisher has developed a Web shop for the newspapers and the brand manager wishes to explore the possibilities for a Web shop with living related products. And finally, commenting is an option where visitors can comment on an article. This feature, which is also available on related websites, could also be

implemented on the brand’s website.

These eight content attributes are believed to have a positive influence on the occurrence of a flow experience and make a website attractive. The hypothesized relations for the individual attributes are presented in Table 1. The following hypotheses are stated:

H1 Content Attributes are positively related to Flow

H2 Content Attributes are positively related to Attractiveness

Table 1 Hypothesized Relations for Content Attributes

Hypothesized Relations Regional content Regional content H1a: + H2a: + Uniqueness User generated content H1b: + H2b: + Own source of content H1c: + H2c: + Third party content H1d: + H2d: + Interactive features Web poll H1e: + H2e: +

Forum H1f: + H2f: +

Web shop H1g: + H2g: + Comments H1h: + H2h: +

Design Attributes

This thesis distinguishes between seven design attributes which are divided into three groups, namely personalization of the website, the presentation style of the content of the website, and the ease of using the website. Websites are not just a static collection of interlinked pages containing only text and some images. Nowadays websites are more dynamic, where interactivity with the website and with its visitors is more important. The following attributes, which is not a exhaustive list, are considered as key elements for a website, such as the interaction with a website, the use of images and videos, and the way the Web pages are interlinked.

Personalization. A website that is personal is for each user a little different, and also empowers the visitor to control

the level of content and design of the website. Empowerment and personalization could increase the customer satisfaction and e-loyalty to the website (Huang et al., 2006). There is made a distinction between four

(10)

profile. Second, empowering the visitor can be achieved through a filter to narrow down or expand the information on a particular subject (to control the level of information detail), and third, to configure the layout and format of the website to fit their own preferences (Huang et al., 2006). Finally, providing customers with customized information can make a website more personal, by sending information about promotions and special offerings that are tailored to each customer’s interests. It is also believed that a personalized website enhances the perceived attractiveness.

Presentation Style. The presentation of the content refers to the amount of multimedia supporting a written text.

There is a distinction between content-based (unanimated), a traditional, standard textual article, and context-based presentation (animated), a video clip of a report alongside a textual summary (Yates and Noyes, 2007). In this thesis, a distinction for the context-based design is made between whether images and videos are used to replace a standard textual article. The outright preference of consumers is a context-based website design (Yates and Noyes, 2007), which was also suggested by Skadberg and Kimmel (2004). This style is more interactive, informal and spontaneous than solely text. As attractiveness and interactivity facilitate the occurrence of a flow experience (Yates and Noyes, 2007), it is therefore believed that the presentation style enhances both the perceived attractiveness of a website and the occurrence of a flow experience.

Navigation Type. It is believed that a website’s ease of use is affected by the type of navigation, that is, the way the

different pages are interlinked. When a website contains a lot of information (not only the amount but also the variation of the information), the website contains usually more pages. Such a website is often perceived as complex (Huang, 2003). The information complexity can be reduced by eliminating the amount of interlinked pages. For example, a clear navigation structure to find information on the website can create an easier use (Huizingh, 2000). As the complexity of a website has a negative impact on the focused attention (Hoffman and Novak, 1996; Huang, 2003), the website’s ease of use is believed to have the reversed impact. It is also believed that a website with a clear navigation structure, where information is easily accessible, is found to be more attractive by visitors.

These seven design attributes are believed to enhance the occurrence of a flow experience and make a website more attractive. The hypothesized relations for the individual attributes are presented in Table 2. The following hypotheses are stated:

H3 Design Attributes are positively related to Flow

H4 Design Attributes are positively related to Attractiveness

Table 2 Hypothesized Relations for Design Attributes

Hypothesized Relations Personalization User profiles H3a: + H4a: + Configuring layout and format H3b: + H4b: + Controlling level of information detail H3c: + H4c: + Providing customized information H3d: + H4d: + Presentation style Replacing test with images H3e: + H4e: + Replacing test with videos H3f: + H4f: + Navigation type Horizontal links H3g: + H4g: +

(11)

2.2 Attractiveness

Part of the central concepts in this thesis is a website’s perceived attractiveness. Finding a website attractive or unattractive is a user’s attitude towards the website; it is one’s internal evaluation of the extent to which a website is favorable (Sicilia and Ruiz, 2007). Although this is an internal assessment of the website’s design and content, it is believed that this is influenced by attributes present on a website.

A website’s perceived attractiveness entails the amount of multimedia used on a website (Huizingh, 2000; Huang et al., 2006). These graphics and animations may take some time to download. However, nowadays this is not such a problem anymore, as 88 percent of the Dutch households has access to the Internet, of which 91 percent has a broadband connection (Centraal Bureau voor de Statistiek, 2007).

The perceived attractiveness of a website is an important contributor to the occurrence of a flow experience, both directly and indirectly (Skadberg and Kimmel, 2004). In this thesis, the visitor’s perception of the attractiveness of a website is believed to be a strong influence on the occurrence of a flow experience on a website. When a website is found to be attractive, one perceived the computer environment as real as the physical environment (Skadberg and Kimmel, 2004). This enhances the occurrence of a flow experience.

It is also believed that attractiveness is also directly related to the intention to return to that website. This is supported by the results of Skadberg and Kimmel (2004), where the attractiveness of a website contributes to taking positive actions such as revisiting to the website, and the results of Yates and Noyes (2007), where attractiveness is recognized as a condition for a website’s effectiveness, i.e. the willingness to return.

2.3 Flow Construct

Despite the vast amount of research based on Csikszentmihalyi’s flow construct, there is no general conceptual model (Ilsever et al., 2007; Siekpe, 2005). The current literature on flow is “too broad and ill defined because of the numerous ways it has been operationalized, tested, and applied” (Koufaris, 2002:203). Therefore, it is important that the definition of flow is deliberately chosen.

The large amount of definitions on flow in a computer environment can be divided into two distinctive ways (Ellis et al., 1994; Shin, 2006). The first regards flow as a function between challenge and skill (e.g. Csikszentmihalyi, 1991). Flow occurs when the degree of challenge and the degree of skills are both equally high. When the degree of challenge is much higher than the degree of skill, one gets frustrated. Contrary, the degree of challenge is much lower than the degree of skill, one gets bored.

Due to the nature of an online computer environment, it is difficult to view flow as the result of the balance between one’s skills and challenges. With the Web being viewed as a multi-activity medium, the traditional measures of the flow state with the comparison of skills and challenges “does not accommodate the unique features of the Web as a multi-activity medium” (Chen et al., 1999: 586-7). The Web is not just navigating and surfing through

hyperlinks. These constructs such as control and challenge are not sufficient, because the “perceived challenges and skills are situated in time and space”, depending on “what it is that the user is doing and the goal(s) that s/he is pursuing” (p.590). Skills and challenges in an online environment are not associated with one’s knowledge of soft and hard ware. Skills should be viewed as “internal, situated concerns of the individual to meet the level of challenge” and challenges are associated with “with the individual's interests and other intellectual sources, such as tracking or retrieving information” (p.602).

Flow as a Mental Mode

(12)

origin of these definitions, which is the work of Csikszentmihalyi (1991, 2000); several authors have used a different definition of a flow experience, consisting of different subconstructs. This led to a broad range of operationalizations of flow, especially on the factors, antecedents and the consequences of a flow experience. Multiple overviews of these definitions can be found in Huang (2006), Ilsever et al. (2007) and Siekpe (2005). In this section, an overview of the mostly used constructs is given.

Webster et al. (1993) investigated the factor structure and correlates of flow in human-computer interactions. In this context, flow “incorporates the extent to which (a) the user perceives a sense of control over the computer interaction, (b) the user perceives that his or her attention is focused on the interaction, (c) the user’s curiosity is aroused during the interaction, and (d) the user finds the interaction intrinsically interesting” (p.413). However, they found that curiosity and intrinsic interest are highly interdependent, and should be labeled cognitive enjoyment. When one’s attention is focused, irrelevant thoughts and perceptions are filtered out, and s/he loses

self-consciousness. One is intrinsic interested when s/he is involved in the activity for its own pleasure and enjoyment (Webster et al., 1993).

Ghani and Deshpande (1994) researched the experience of individuals using computers in the workplace, in relation to job characteristics. In their model, flow is characterized by the intense concentration in an activity and the enjoyment one derives from an activity. Flow is determined by the individual's sense of being in control and the level of challenge perceived in using computers. Their results indicated that for individuals holding jobs with high variety, identity, autonomy, and feedback, perceived control played a greater role. For low task-scope jobs, challenge was more important. Flow was found to link significantly with exploratory use behavior, which was linked to extent of computer use. Also, there was a direct significant and positive relation from perceived control to exploratory use.

Hoffman and Novak (1996) were the first to test flow in a computer-mediated environment (CME). They defined a flow experience in a CME as “the state occurring during network navigation, which is (1) characterized by a seamless sequence of responses facilitated by machine interactivity, (2) intrinsically enjoyable, (3) accompanied by a loss of self-consciousness, and (4) self-reinforcing” (p.57). Their process model, Hoffman and Novak (1996) stated fifteen research issues flow in a hypermedia computer-mediated context, to describe online consumer experiences.

Based on this conceptual model, Novak et al. (2000) developed a structural model of flow in the context of online shopping and Web application usage. They included 1600 Web users in their sample on flow experience, measured by 13 constructs. They defined flow as a “cognitive state experienced during navigation that is determined by (1) high levels of skill and control; (2) high levels of challenge and arousal; and (3) focused attention; and (4) is enhanced by interactivity and telepresence” (p.22). However, they failed to operationalize their concepts, such as flow, skill and challenge, in terms of specific Web activities (Chen, 1999; Pace, 2004). Also, these thirteen subconstructs are not exhaustive (Finneran and Zhang, 2005).

In their research, Skadberg and Kimmel (2004) evaluated visitors’ experience while browsing a website with regard to the flow concept in the context of human-computer interactions. Flow is indicated by time distortion (defined as focused attention and involvement), enjoyment and telepresence. They proposed that flow was influenced by the website’s design, contents and performance, and the visitor’s prior Web experience and knowledge.

Representations of the Flow Construct

(13)

antecedents influencing the flow experience (e.g. Ghani and Deshpande, 1994; Hoffman and Novak, 1996; Novak et al., 2000). In this model, the direction of causality is from the first-order factors to the flow concept.

The subconstructs mentioned in the preceding part are represented in Table 3. These are labeled either as antecedents, which determine flow, or as factors, which characterize flow. Siekpe (2005) found that of the mostly used constructs, there was only agreement on two constructs (i.e. enjoyment and telepresence). He tested the remaining first-order (FO) variables, which can be identified in Table 3, both as factors and as antecedents, and concluded that the reflective model has a slightly higher explanatory power.

This has three implications for the assessment of flow, namely “(1) the direction of causality is from the higher-order (HO) construct to the FO (first-higher-order) latent variables, (2) the FO variables are manifestations of the HO construct, and while changes in the FO variables should not cause changes in the HO construct, changes in the HO construct do cause changes in the FO variables, (3) the FO variables should share a common theme and are expected to covary with each other” (Siekpe, 2005:40-1).

These subconstructs are highly interdependent and are based on different interpretations of the work of Csikszentmihalyi (1991, 2000). In this thesis, the flow model of Webster et al. (1993), one of the most frequently mentioned models, is adopted. This model has the advantage of consisting of only factors instead of antecedents. Also, the reliability of this model has been proven in literature. Despite the model originating in 1993, far before the use of the Internet in daily life, browsing websites is an activity where flow can occur (Chen et al., 1999). The factors are described in the next section.

Table 3 Subconstructs of Flow, in Literature Identified as Either Antecedents or Factors Ghani and

Deshpande (1994)

Hoffman and Novak

(1996) Koufaris (2002) Novak et al.(2000) Siekpe (2005)

Skadberg and Kimmel

(2004) Webster et al.(1993) Enjoyment Factora Factorb Factorb

Intrinsic interest Factora

Curiosity Factorb Factor

Focused attention Antecedentc Antecedentc Factorb Factorb Factora Concentration Factora Factor

b

Control Antecedenta Consequencec Factorb Factorb Factora

Skill Antecedentc Antecedent

c

Antecedenta Challenge Antecedenta Antecedentc Antecedentc Antecedentc Factorb Antecedenta Telepresence Antecedentc Antecedentc Antecedentc Note: Originated from a Csikszentmihalyi (1991), b Ghani and Deshpande (1994), c Hoffman and Novak (1996)

2.4 Flow Factors

In this thesis, flow is considered to be a multidimensional construct reflected by four first-order variables, namely intrinsic interest, curiosity, focused attention, and control. Also Huang (2003), Huang (2006), and Siekpe (2005) consider these as variables. These flow factors are represented in Figure 1.

Curiosity. Curiosity and intrinsic interest were the two factors that Webster et al. (1993) merged together because

they found evidence that both subconstructs were highly interdependent. “In computer interactions, flow involves a certain kind of intrinsic interest. This is not the intrinsic interest that arises from a game of chance, for example, but rather the intrinsic interest that accompanies cognitive arousal and use of the imagination” (p.420). However, in other research, the merger of curiosity and intrinsic interest was not always recognized.

(14)

However, Huang found that curiosity replaced attention focus with regard to flow and situational involvement. An explanation for this replacement is that curiosity reflects the internal state of flow, and attention focus reflects the outputs of flow. “Curiosity is seen as a motivational variable that represents desire for information” (p.404) and “attention is a series of activities in which consumers selectively allocate cognitive resources” (p.405).

Figure 1 Flow Construct and the First-Order Factors

Intrinsic Interest. When one is in a state of flow, one experiences a joyful interaction with the technology

(Csikszentmihalyi, 1991). Enjoyment was found to characterize flow by Csikszentmihalyi (1991) and Ghani and Deshpande (1994). Webster et al. (1993) regarded intrinsic interest as a factor of flow. However, the measurement instruments of enjoyment by Ghani and Deshpande (1994) and intrinsic interest by Webster et al. (1993) are almost identical. According to (Csikszentmihalyi, 1991), enjoyment and intrinsic interest are two components of the complex variable ‘enjoyment’. Siekpe (2005), in his research on the multidimensionality of flow, also merged these two subconstructs into ‘enjoyment’. However, because there was agreement in literature that enjoyment should be regarded to as a factor, Siekpe (2005) did not measure this subconstruct.

Focused Attention. When one is in a flow state, one experiences that his attention is completely absorbed by the

activity to the extent that nothing else matter (Csikszentmihalyi, 1991). This factor is named ‘concentration’ (e.g. Chen et al., 1999; Ghani and Deshpande, 1994; Hoffman and Novak, 1996; Koufaris, 2002) or ‘focused attention’ (e.g. Hoffman and Novak, 1996; Novak et al., 2000; Koufaris, 2002; Webster et al., 1993). The mostly used definitions are from Webster et al. (1993) and Ghani and Deshpande (1994). Webster et al. (1993) defined focused attention as “the user perceives that his or her attention is focused on the interaction” (p.413). Also the four-item scale of Ghani and Deshpande (1994) is frequently used, measuring concentration as absorption, focused attention, concentration, and engrossment. In his research on the multidimensionality of flow, Siekpe (2005) merged these two subconstructs into ‘concentration’.

Control. In his research on online flow symptoms, Chen (2006) excluded the dimension of matched challenges and

skills because of “its difficulty in conceptually measuring a Web users’ challenges and skills in distance without observing on site and tap into his/her inner experience first” (p.226). In this thesis, the definition of flow as the function of skill and control is not adopted. Like the view of Chen (2006), this thesis excludes skill and challenge but includes perceived control as a factor characterizing flow.

There is a lack of consistency over the subconstruct control (Huang, 2006). Perceived control has been proposed as an antecedent, a factor or a consequence of flow. Control as an antecedent is defined as the individual's perception that s/he exercises control over the interaction with the technology (e.g. Ghani and Deshpande, 1994). As a factor, flow is defined as the user’s perceived sense of control over the computer interaction (e.g. Csikszentmihalyi, 2000; Webster et al., 1993). Finally, as a consequence of flow, Hoffman and Novak (1996) define ‘perceived behavioral control’ as the perceived ease or difficulty of performing the behavior.

Control was one of Siekpe’s (2005) subconstructs for determining whether flow should be regarded as a reflective or formative model. He used Ghani and Deshpande’s (1994) conceptualization of control as an antecedent. As an

Focused attention Control Curiosity

Flow Experience

(15)

antecedent of flow, ‘control’ was found not statistically significant. Siekpe (2005) therefore concluded that ‘control’ should be regarded as a factor. Also Huang (2006) regarded control as a factor.

These four subconstructs of flow – curiosity, intrinsic interest, focused attention and control – are regarded as characteristics of flow. These subconstructs are not inexhaustible, as is proven by the large amount of discussion on the representation of the flow construct. However, in literature it is believed that these subconstructs are

characteristics of flow. As stated before, one of the positive consequences of a flow experience on an attractive website is that one wants to recreate the state of flow.

2.5 Intention to Return

Flow is said to generate an intention to return to the website, as proposed in Ilsever et al. (2007), Koufaris (2002), and Siekpe (2005). In their flow model when browsing a tourism website, Skadberg and Kimmel (2004) propose that flow leads to attitudinal and behavioral change in a visitor (i.e. inquiring for more information after visiting the Website, coming back to the website later for information, and taking positive actions, such as visiting the place presented).

Koufaris et al. (2001) and Koufaris (2002) researched the flow model in an e-commerce context and whether this would lead to intention to return and intention to purchase from the website. However, they found no relation to intention to purchase, but they did find a relationship with the intention to return.

Choi and Kim (2004) researched which design factors are important for customer loyalty in the context of online games. Their results confirmed their hypothesis that “the features of efficient interaction in online games can indeed influence customer loyalty through the experience of flow” (p.18).

In Palmer’s (2002) research on corporate websites, for the website’s navigation (i.e. arrangement, sequence, links, and layout,) and content (amount and variety of information, and content quality), there was a significant relationship with website success (defined by the subconstructs ‘user satisfaction’, ‘likelihood of return’, and ‘frequency of use’).

The previous three concepts are central in this thesis. It is believed that website which is perceived as attractive will be more likely to enhance the occurrence of a flow experience on that website, and that an attractive website will increase the intention to return to a website. Also, it is believed that a website where the joyful experience of a state of flow is experienced is more likely to be visited again. Therefore, the following hypotheses are stated:

H5 Attractiveness is positively related to Flow

H6 Attractiveness is positively related to Intention to Return H7 Flow is positively related to Intention to Return

This leads to the following conceptual model, presented in Figure 2. The stated hypotheses are also presented.

Figure 2 Conceptual Model

(16)

3. METHODOLOGY

3.1 Case Description

The brand, as mentioned before, consists of a website, two newspaper supplements, a (recently cancelled) glossy magazine and a fair. The content of the supplements is nearly the same, and the main difference is the amount of advertising and the title of the newspaper. In two relatively recent market researches on the supplement (in June 2007) and the magazine (in June 2008), the brand’s target group consists of:

• Females who are middle aged (families with children and middle aged couples without children);

• Who have a high income (modal and twice modal) or double-income couple;

• And a high education level (secondary education and higher education);

• Who own a house (mostly free-standing), with a garden; and

• Live in the capitals of the three northern provinces of the Netherlands.

3.2 Study Design

In previous literature on flow in an online context, flow was measured by self-reports in two ways. The first method is the experience sampling method (Huang, 2006), where participants are signaled at random times throughout the day to complete a self-report form to measure the immediate conscious experience. However, there are some issues regarding this method for examining flow, such as its measures are too simple for the complex variables of challenges and skills and that it does not take into account the effects of individual differences (Huang, 2006).

The second method is self-report flow scales, which has two variations, is used “to measure flow so as to capture the subjective state while minimizing interference” (Huang, 2006: 393). The first variation is that respondents provide personal examples of flow events based on brief descriptions of flow events given, rate the overall flow experienced while using the Web (e.g. Novak et al., 2000). The second variation measures flow using Likert scale statements (e.g. Webster et al., 1993) or bipolar semantic-differential scales (e.g. Ghani and Deshpande, 1994).

This thesis adopts the second method, where the flow experience on a specific website is measured by different Likert-scale statements. Such statements are used for a couple of reasons. First, Likert-scales are appropriate for studying subjective states such as flow (Webster et al., 1993). Second, because of the multi-dimensionality of the flow construct, the statements in the measures of flow and intention to return are all based on existing literature. Third, these scales are used in a majority of the research on flow and were found to be valid.

3.3 Data Collection

This research is split into two studies. Respondents’ interests in website attributes and the flow experience is measured in study 1, by means of a questionnaire (see Appendix A). The interests and the extent to which these interests correspond with the results are discussed later in this thesis. In the questionnaire, respondents are asked for the URL (Uniform Resource Locator; the Web address of the website) of a website they visit often and for a long period of time, which is comparable to the research of Huang (2003). The website mentioned by a respondent is also used as the basis for the questions regarding the flow factors. These websites are the basis for study 2, where the different content and design attributes are measured to access which attributes lead to a flow experience and the respondent’s perceived attractiveness of a website. Both studies are described in the remainder of this section.

Study 1. The questionnaire which forms the basis for study 1 is presented to members of a panel and visitors of the

(17)

elderly people are overrepresented, a selection criterion was inserted. The brand’s target group is females between 25 and 55 years old. For comparison purposes are also males included.

The questionnaire was placed on the panel administrator’s server. A notification was sent out to the panel members. The questionnaire was available on Thursday 4 September, and was closed 13 days later. After one week, a reminder was sent out to the panel’s members. On the brand’s website a hyperlink was placed with the notification of the incentive (five free magazines of a related brand, due to the decision that the website brand’s magazine was cancelled). A newsletter with the same notification was sent out to 2100 email addresses from visitors of the previous brand fair. An editorial notification of the questionnaire with a reference to the hyperlink on the website was

published in the brand’s supplements. In the second week, the same notification was published as a reminder. A total of 1,732 respondents filled in the questionnaire. Half of the cases (863) were discarded for not filling in a website which is frequently visited and therefore not answering the flow related questions. The remaining 869 cases were analyzed for invalid responses. The criteria for this analysis were multiple entries (as the question stated to fill in only one website), general terms such as “dating sites” or “news”, non-response such as “none” or “multiple interests”, or websites such as search engines or web mail accounts. When there was doubt on whether or not the website was an actual URL (for example, when the entry was solely “nu” instead of “nu.nl”), the entry was entered into search engine Google. When the entry was not located on the first page, whether or not in the title of the website or in the URL, the entry was recognized as invalid. Based on these criteria, 202 cases were discarded.

The remaining 667 respondents list 247 unique websites, of which seventeen are mentioned over five times by 395 respondents, as presented in Appendix B. This leads to a set of 329 cases, which consists of 45.3 percent males and 54.7 percent females. The distribution of this sample into the seventeen unique websites is also presented in Appendix B. These seventeen websites are analyzed in study 2.

Study 2. The basis for study 2 is the websites that are visited frequently and for a long period of time. It is believed

that a few websites are mentioned often and a lot of websites are mentioned a couple of times. Therefore, of the valid responses, only those seventeen unique websites mentioned by a larger number of respondents are suitable for further analysis. These websites are analyzed based on certain content and design attributes.

This study is carried out because adding these attributes to the questionnaire would make the questionnaire severely lengthy and are not so much related to the repositioning of the brand. Also, people do not precisely recall the website and its attributes (Chen, 2006; Finneran and Zhang, 2005).

To undo this subjectivity, the researcher will carry out the analysis of these website attributes. However, this is also relatively subjective. Standardization of this study will reduce the subjectivity of the researcher. Also, to enhance the objectivity, the different website attributes are measured according to the presence of an item. These attributes are mentioned in the next section.

This study was carried out by creating a table with the measures horizontally and the URLs of the websites vertically. Each website was individually analyzed and scored. When the item is present on a website, the researcher notes this by adding a ‘1’ into the corresponding cell. This is done twice, in case the researcher overlooked the attribute the first time. These scores are then compared. If a score does not match the previous score, a third person is asked to give her opinion. For this study, the main content of the websites are regarded – content of ‘about us’ or ‘help’ pages were not analyzed. These scores were later merges with that of study 1.

3.4 Measures

(18)

Content Attributes

These measures consider whether a content attribute can create a flow experience. These attributes are measured in study 2. The following measures of content attributes are hypothesized to influence the occurrence of a flow experience and the perceived attractiveness of a website:

Regional Content. This item reflects the regional scope of the content. However, due to the nature of the theme, the

content of the website needs to be more (inter)national focused (for example design products). The interests of the respondents in three different content types with a regional view were measured in study 1. The presence of Regional

content is measured in study 2 by an option that enables the content to be filtered on a geographical area.

Uniqueness. The uniqueness of the content refers to the source of the content. In this thesis, three content sources

are differentiated, namely content created by the website’s editors, by third parties or by the website’s users. Also, unique content can originate from users, also known as user generated content. The intention of the visitor to participate in the content is measured in study 1.

In study 2, the websites are scored on the source of the website, namely Own source of content (generated by editors of the website), Third party content (generated by a third party), or User generated content (content generated by users). This could be indicated on a website, for example when a reference is made to the author of an article or the original source. When there is reason to assume third party content (for example from press agencies), a Google search was carried out to access whether this content is also published on other websites.

Interactive Features. In this thesis, four types of features are included in the questionnaire, namely a Web poll, a

forum, a Web shop and the option of commenting on an article. In study 1, the respondent’s interests for the first threw features are measured. In study 2, the presence of all four interactive features is measured.

First, the presence of a Web poll is measured by the presence of statements with two or more clickable options. A

Web poll is a tool where users’ opinions are collected and graphical reproduced so that the representation of the opinions is visible. Second, the presence of a Forum is measured by the presence of a ‘forum’ or ‘discussion board’, where users can discuss topics in threads with other users. Third, the presence of a Web shop is measured by whether additional physical or digital goods or services can be purchased via a financial transaction with the website, that that are not necessary for the use of the website (i.e. paying for the use of placing content on the website). Finally, the presence of Commenting on a page is measured by the option for users/visitors to leave a comment on/to respond to the topic on the page.

Design Attributes

The following measures of design attributes are hypothesized to influence the occurrence of a flow experience and the perceived attractiveness of a website:

Personalization. Personalization refers to the extent to which a user/visitor can adjust the content and design of the

website to satisfy his/her own interests. In study 1, the degree of the interest in personalization is measured by the extent to which the layout and format of the website can be configured to fit their own preferences and to which the level of information detail can be controlled by the visitor (Huang et al., 2006).

(19)

adjusted, via the ‘help’ pages or a guest account the preferences in the user account. Third, the presence of Controlling

the level of information detail refers to an option for excluding certain panels or topics on the website, whether via the user profile or via cookies. Finally, the presence of Proving the user with customized information refers to when the user is notified with (the status of) personal interests, whether via an e-mail or as a notification on the website at each visit.

Presentation Style. The presentation of the content refers to the extent in which only text is used to present the

information. According to Skadberg and Kimmel (2004) and Yates and Noyes (2007) there are two presentation styles, context and content-based designs. In this thesis, there is a distinction between images and videos replacing a standard textual article. In study 1, this is measured by the respondent’s preference of watching videos in stead of reading text. In study 2, this is measured by the presence of whether images and graphics, and video files are used to replace text.

Navigation Type. A website’s ease of use refers to the navigation structure of the website (e.g. Huizingh, 2000). In

study 2, this is measured by the navigation type. Huizingh (2000) distinguished four types of navigation structure. One extreme is the tree or hierarchical structure, where a page is linked to one page on the previous level and one or more pages at the next level. The other extreme is the extensive network, where a page is linked to all other pages. This is creates a very complex website with many paths. Therefore, there are two more structures with a limited number of paths, namely a tree with a return-to-home page button and a tree with a few horizontal links. The graphical presentation of these four types is presented in Figure 3, where the squares represent a content page and the arrows represent a hyperlink from one page to another.

Hierarchical structure (tree) Tree with home-page button Tree with few horizontal links Extensive network

Figure 3 Navigation Types

Attractiveness, Flow Experience, and Intention to Return

The measures of perceived attractiveness, flow experience and the intention to return are stated on a five-point Likert scale, ranging from strongly disagree (1) to strongly agree (5). The scales have been adjusted to the specific topic of this thesis (e.g. ‘the Web’ was substituted by ‘your previously mentioned website’). The original measures for flow (see Appendix C) were stated on a seven-point scale, but due to third party restrictions, only scales with five points were permitted.

Attractiveness. The perceived attractiveness of the website refers to the extent one finds the graphic representation

of the website’s content attractive. This attribute is used to access whether the regularly visited website is attractive and if this has an impact on flow. This item is measured in study 1 by the degree of attractiveness of the frequently visited website, mentioned by the respondent previously in the questionnaire.

(20)

Flow Experience. In this thesis, flow is differentiated into four dimensions: curiosity, intrinsic interest, focused attention, and control. The two mostly used measures of flow factors originate from Webster et al. (1993) and Ghani and Deshpande (1994). Both these scales show good reliability, of .82 and an average of .818, respectively. The first three dimensions are based on the measures of Webster et al. (1993), because this model views the flow dimensions as factors, as Ghani and Deshpande (1994) view two dimensions as factors and two as antecedents. This thesis adopts the Webster et al. (1993) representation as factors.

Curiosity refers to the extent to which the user’s curiosity is aroused during the interaction with the website (Webster et al., 1993). This three item scale for curiosity of Webster et al. (1993) is found to be reliable by Huang (2003:433) and Huang (2006:399), with a scale reliability of .71 and .86, respectively. The average variance extracted was 45.2 percent and 67 percent, respectively. Intrinsic interest refers to the extent to which the user finds the interaction with the website intrinsically interesting (Webster et al., 1993). This three item scale for intrinsic interest of Webster et al. (1993) is found to be reliable by Huang (2003:433), with a scale reliability of .72 and an average variance extracted of 48.3 percent. Focused attention refers to the extent to which the user perceives that his or her attention is focused on the interaction with the website (Webster et al., 1993). This three item scale for focused attention of Webster et al. (1993) is found to be reliable by Huang (2003:433) and Huang (2006:399), with a scale reliability of .82 and .75, respectively. The average variance extracted was 60.6 percent and 51 percent, respectively.

Control refers to the extent to which the user perceives a sense of control over the interaction with the website (Ghani and Deshpande, 1994). The Webster et al. control scale was found to have a moderate reliability (Huang, 2003; Huang, 2006). Also, Webster et al. (1993) and Huang (2003) excluded one item on control due to unreliability. Therefore, the fourth dimension is based on the control scale by Ghani and Deshpande (1994). This three-item control scale shows more reliable results in Koufaris (2002:215) and in Siekpe (2005:39), with a scale reliability of .813 and .90, respectively. Due to the nature of this control scale (semantic differential), this scale is adjusted into a five-point Likert scale.

As the flow construct is based on four factors, a factor analysis is carried out in order to access which items load onto a factor. The twelve items were analyzed using an oblique rotation, because it is believed that the underlying factors are related due to the multidimensional nature of the flow construct. The factor analysis shows that three components have an eigenvalue above 1 and are therefore distinguished. Four items were dropped and every remaining item load onto one factor.

The components are presented in Table 4, where only correlations above .60 are stated. It shows that the factors Curiosity and Intrinsic interest are regarded as one factor. This is in line with the results of Webster et al. (1993), where the combined factor was regarded as ‘Cognitive enjoyment’. The KMO statistic is .639, which is an acceptable but mediocre value. The Bartlett’s test, to test whether there are significant correlations between the variables, is significant (p < .00), and therefore the factor analysis is appropriate.

Following the factor analysis, flow is measured by the eight remaining items (dropping one item for each factor) and is regarded as one variable. The scale is found to be moderately reliable, with a Cronbach’s alpha of 0.646. This is less than found in previous literature, as stated above. However, it is sufficient enough to regard to flow as one construct.

Intention to Return. In order to access the consequence of a flow experience on a website, the respondent’s

(21)

Table 4 Pattern Matrix for Flow Factors in Factor Analysis

Cognitive

enjoyment attention Focused Control CUR1 Visiting this website excited my curiosity. .865

CUR2 Interacting with this website made me curious. .861 CUR3 Visiting this website aroused my imagination. - INT1 Visiting this website bored me. (R) - INT2 Visiting this website was intrinsically interesting. .730 INT3 This website was fun for me to use. .677

FAT1 When visiting this website, I thought about other things. (R) .847 FAT2 When visiting this website, I was aware of distractions. (R) .816 FAT3 When visiting this website, I was totally absorbed in what I was doing. - CTR1 I clearly knew the right things to do when visiting this website -

CTR2 I felt calm when I visited this website. .916 CTR3 When visiting this website, I felt in control over this website. .895 Extraction Method: Principal Component. Rotation Method: Oblimin with Kaizer Normalization. Rotation converged in 6 iterations. Note: CUR = Curiosity, INT = Intrinsically interesting, FAT = Focused attention, CTR = Control.

Control Measures. The control measures are Gender and Internet use. According to Shin (2006), gender was not

influential in determining the level of flow, but the female participants show to have a lower perception of their skills and challenges level. However, as the level of skills and challenges are not a part of this thesis, gender is regarded a control measure. Due to the repositioning of the brand’s website, the outcome of whether males or females are more susceptible to a flow experience could be relevant.

(22)

4. ANALYSIS

4.1 Descriptive Statistics

The descriptive statistics for the variables in this thesis are presented in Table 5. Variables 1-3 correspond to the variables in the central hypotheses. The groups 4-11 and 12-18 represent the content attributes and the design attributes respectively. Variables 19 and 20 stand for the control variables. The scores for the content and design attributes are presented in Appendix D. The correlation coefficients are presented in Appendix E. Furthermore, two notes must be made regarding two variables. First, for the design attribute Navigation type, the measure of Horizontal links was inserted, as the two extremes of the navigation types are not scored. And second, the control variable Internet is normalized using a logarithmic transformation, due to right skewness.

The variables are tested for violation against the assumptions for normality, linearity, homogeneity and multicollinearity. To test for the normality, the skewness and kurtosis statistics are presented in Table 5. The distributions of all variables were (slightly) skewed and peaked. The statistics for Intention to return, Flow, Attractiveness and the transformed control variable Internet use all fall in the range of ± 1.0, and are therefore considered to have an acceptable normal distribution. To test for the linearity of the variables, the correlations between the dependent variable Intention to return and the independent variables are shown in the first column in Appendix E, labeled ‘1’. The correlations with Intention to return is significant for Flow, Attractiveness, and the control measure Internet use. Testing for linearity is not possible for Flow and Attractiveness, as the content and design attributes are nominal variables. To test for the homogeneity of the variances, the Levene statistic was measured, as presented in Table 5. For all variables, this statistic was not significant and these are therefore regarded as homogeneous. Therefore, the assumption of homoscedasticity is not violated. Finally, the model is tested for multicollinearity, which is indicated by scores above 5 on the variance inflation factor (VIF), as presented in Table 5. The variables Attractiveness, Flow and the control variable Internet use have a VIF score below 5. This indicates that multicollinearity is not a problem.

Table 5 Descriptive Statistics (n=329)

(23)

4.2 Results

Content Attributes

In the remainder of this section, the results of the regression analyses are described. Table 6 shows the regression analysis for flow with the content attributes. Model 1 is significant and explains 5.6 percent of the extracted variance. Seven content attributes show a significant relation with a flow experience. First, the impact of Regional content is highly significant, but has a negative relationship. This is supported by the correlation between flow and Regional

content. Second, it also shows that the uniqueness of the content is negatively related to flow, as User generated content and Own source of content are negatively related to a flow experience. Third party content shows no significant relation to flow. The two significant unique content characteristics are not correlated with the flow experience. However, there is a significant correlation between User generated content and two flow factors, which are significantly correlated with the flow construct. Own source of content is not significantly correlated with neither flow nor a flow factor, but it is significantly correlated with Regional content. Finally, all interactive features show a significant relationship with flow. The presence of a Web poll, Forum and Comments is positively and (highly) related, but the presence of a Web shop negatively. None of these interactive features show a correlation with the flow construct. However, these features, except for Forum, are correlated with different flow factors. Also, these features are significantly correlated with each other, and are significantly correlated with Regional content.

The regression analysis for attractiveness with the content attributes is presented in Table 7. Model 2 is significant and explains 7.1 percent of the extracted variance. Three content attributes show to have a significant impact on the perceived attractiveness of a website, and are also related to the occurrence of a flow experience. First, the presence of Regional content has also a significant impact on the website’s attractiveness, but is again negative. This is supported by the negative correlation with attractiveness. Second, as for the unique content characteristics, only the presence of

User generated content has a significant but negative relation with the attractiveness of a website. This is also supported by the negative correlation with attractiveness. Third, only one interactive feature, namely the presence of Comments has a positive significant impact on attractiveness. However, this is not supported by a correlation with attractiveness, but it is correlated with other interactive features that are correlated to attractiveness.

Design Attributes

The regression analysis for the flow with the design attributes is shown in Table 8. Model 3 is slightly significant and explains 4.2 percent of the extracted variance. Three design attributes show a significant relationship with a flow experience. First, the personalization characteristic Providing customized information is positively related to flow. This variable is not directly correlated with flow and also not with the flow factors. Finally, both presentation styles show a (highly) significant but negative relation with flow. This is not supported by a significant correlation with flow. However, Replacing text with images is correlated with the flow factor Focused attention and Replacing text with videos is correlated with the factor Cognitive enjoyment, and both flow factors are significantly correlated with the flow construct.

Which design attributes relate to the perceived attractiveness of a website is presented in Table 9. With an explained variance of 8.5 percent, is Model 4 significant. Three design attributes show a significant relation with the perceived attractiveness of a website, of which only Providing customized information is also related to flow. First, the personalization characteristic User profiles has a highly significant negative impact, which is also supported by the negative correlation with attractiveness. Second, Providing customized information, another personalization

(24)

Table 6 Regression analysis for Flow Experience with Content Attributes (n=329)

Table 7 Regression analysis for Attractiveness with Content Attributes (n=329)

Baseline Model 1 Baseline Model 2 Constant 3.780 ** (.121) 3.997** (.190) Constant 4.100 ** (.190) 4.368** (.298)

Gender .002 (.057) .015 (.058) Gender -.117 † (.089) -.092 (.091)

Internet use .024 (.036) .031 (.037) Internet use .050 (.057) .052 (.058) Regional content -.314** (.110) Regional content -.289* (.172)

User generated content -.204† (.126) User generated content -.319(.198)

Own source of content -.236† (.152) Own source of content -.027 (.237)

Third party content .023 (.092) Third party content .082 (.143) Web poll .290** (.120) Web poll -.125 (.187)

Forum .271** (.116) Forum .105 (.182)

Web shop -.244* (.135) Web shop -.159 (.211)

Comments .231* (.115) Comments .272(.181)

R2 0.1% 5.6% R2 0.8% 7.1%

F .214 1.874* F 1.300 2.413**

Table 8 Regression analysis for Flow Experience with Design Attributes (n=329)

Table 9 Regression analysis for Attractiveness with Design Attributes (n=329)

Baseline Model 3 Baseline Model 4 Constant 3.780 ** (.121) 4.070** (.236) Constant 4.100 ** (.190) 4.962** (.364)

Gender .002 (.057) .007 (.059) Gender -.117 † (.089) -.109 (.091)

Internet use .024 (.036) .026 (.037) Internet use .050 (.057) .046 (.058) User profiles -.062 (.116) User profiles -.557** (.179)

Configuring layout and format -.061 (.099) Configuring layout and format .078 (.153) Controlling level of information detail -.130 (.118) Controlling level of information detail -.027 (.182) Providing customized information .155* (.084) Providing customized information .273* (.130)

Replacing text with images -.236* (.137) Replacing text with images .057 (.211)

Replacing text with videos -.202** (.086) Replacing text with videos .057 (.132)

Navigation type -.022 (.177) Navigation type -.599* (.273)

R2 0.1% 4.2% R2 0.8% 8.5%

F .214 1.548† F 1.300 3.283**

Table 10 Regression analysis for Flow Experience with

Attractiveness (n=329) Table 11 Regression analysis for Intention to Return with Attractiveness (n=329) Baseline Model 5 Baseline Model 6 Constant 3.780 ** (.121) 2.703** (.172) Constant 2.539 ** (.267) 1.658** (.412)

Gender .002 (.057) .033 (.052) Gender .020 (.125) .045 (.124) Internet use .024 (.036) .011 (.033) Internet use .155 * (.080) .144* (.080)

Attractiveness .263** (.032) Attractiveness .215** (.077)

R2 0.1% 17.2% R2 1.1% 3.4%

F .214 22.428** F 1.8483.851**

Table 12 Regression analysis for Intention to Return with Flow Experience (n=329) Baseline Model 7 Constant 2.539 ** (.267) .586** (.520) Gender .020 (.125) .019 (.122) Internet use .155 * (.080) .142* (.078) Flow .517** (.119) R2 1.1% 6.5% † p < 0.1, * p < 0.05, ** p < 0.01 (one-tailed)

Referenties

GERELATEERDE DOCUMENTEN

The EPP demands a determined application of the new instruments which have been developed in the framework of Common Foreign and Security Policy (CFSP), among which are recourse

The present work resulted in the estimation of a set of ten financial indicators, obtained through the competing methodologies of principal component analysis applied according

One can say that postcolonialism and decolonising campaigns cannot be separated from theological schemata of interpretation that portrayed God in images that represented Western

Consumers who try a brand product perceive a more pronounced brand personality, form a more favorable brand attitude and perceive a stronger brand activation

In addition, we therefore analyzed the effects a more hedonic brand attitude has on the individual components of Customer Performance, which showed that a brand store with a

This study showed that visitors are not having any trouble with a substantial amount of verbal content regarding the usability of the website and they profit from it by a

By focusing on individuals’ need for self-reflection, need for cognition, social comparison orientation and degree of similarities between gossip receiver and gossip target,

The description of the justice of the peace in Belgium shows, unlike in France, a lively practice, which can serve as an example for the Dutch judicial system.. The