Causal Attributions and Expectancy Estimates of Commercial Web Surfers
 

Abstract
How will surfers respond to their success or failure in searching commercial Web sites?
To which factors do they attribute Web searching performance? How would a performance affect subsequent expectancy estimates, such as continuing consideration of Web sites as a commercial purchase or inquiry tool? To answer these questions, this study researches consumer behavior in searching commercial Web sites. Attribution theory and three flow models of network navigation in Computer-Mediated Environments derived useful psychological factors from Web surfers. Two studies are also executed to verify the extracted factors. Finally, the study will scrutinize the relationships between the attribution dimensions and surfersí expectancy estimates toward future Web surfing.
 

 1. Introduction
 Doing what other mass media fail to do, the World Wide Web (WWW) is emerging as one of the most popular mass media in the late 20th century. Technological capabilities, such as e-mail, online chat rooms and video- conferencing, develop personal interactivity. The reciprocal addressability of both information sources and surfers (Web users) makes individualized point-to-point communication possible in the context of mass communication (Eighmey, 1997, p.59).
 The interactive capability of the WWW is considered a versatile vehicle not only to exchange general information but also to deliver commercial advertisements. As a commercial medium, the Web has the ability to facilitate global sharing of information and resources. It has the potential to provide an efficient channel for advertising, purchasing, and even direct distribution of certain goods and services. One study (Bare, 1996) found more than half of the nationís 200 most heavily advertised brands were represented with Web sites. Advertising expenditures on the Internet rose to $351 million in the first quarter of 1998, a 271 percent increase over spending in the first quarter of 1997 (Internet Advertising Bureau, 1998). Unquestionably, Web sites present advertisers with opportunities and challenges, including the need for more systematic research. Advertising and marketing practitioners will benefit from understanding how surfers perceive the Web as a source of advertising (Alwitt and Prabhaker, 1994; Becker, Martino, and Towners, 1976; MacKenzie and Lutz, 1989). Academics are aware also that more systematic study is required to reveal the true nature of commerce on the Web (Berthon, Pitt, and Watson, 1996).
 While surfing the Web, we are fascinated by its speed and convenience to deliver purchases or inquiries. However, frequently we fail to get the commercial information that we want to find, despite consuming lots of time.
 How will surfers respond to their success or failure in searching commercial web sites?  To which factors do they attribute Web searching performance? How would a performance affect subsequent expectancy estimates, such as continuing consideration of Web sites as a commercial purchase or inquiry tool?
 To answer these questions, this study will research consumer behavior in searching  commercial Web sites. Researchers suggest that attribution theory (Kelly 1967, 1972) provides a useful framework in the study of consumer behavior (Mizerski, Golden, and Kernan 1979; Gorn and Weinberg 1984; Sparkman and Locander 1980).
 This study will focus on three major estimates. First, the Web surfersí attribution factors toward Web searching performance will be found. The previous factors of flow state in Computer-Mediated Environments (Hoffman and Novakís conceptual model, 1996; Novak, Hoffman and Yungís causal model, 1997) will be assigned as two attribution dimensions: locus and stability (Weiner 1985, 1986 and Teas and McElroy, 1986). These selected variables in a CME will be re-measured as surfersí attribution factors toward their performance
 Two studies will be executed to verify the extracted variables. The first study will collect surfersí attribution items by asking why they failed or succeeded in finding the commercial information that they were looking for. Using factor analysis, the second study will group these attribution items to derive surfersí attribution factors. The attribution factors will be compared with the selected variables in network navigation of a CME. Second, these attribution factors will be designated by two causal dimensions (locus and stability) and verified by Web users. For example, the interactivity of computer at a given time during the Activity/Survey will be allocated to the external unstable dimension and the interactivity of overall computers at a general Computer-Mediated Environment will belong to the external stable division. Finally, the study will scrutinize the relationships between the attribution dimensions (internal stable, internal unstable, external stable, and external unstable) by surfersí performance level (high and low) and their expectancy estimates toward future Web surfing.

 2. The necessity of this study
 The World Wide Web has attracted a great deal of attention in recent years not only in commercial business but also popular culture. This study, which examine surfersí causal attributions and expectancy based upon latent variables in a process of network navigation, has the following significance.
 First, the main purpose of this study is to test surfersí performance-attribution-expectancy linkages during their Web searching. To understand surfersí attributions toward the performance level of their commercial Web searching is very important for making firm marketing strategies. For example, surfersí negative expectancy derived from their internal/external attributions will be considered useful information to understand consumer behavior and expectancy.
 Second, latent variables in a process of network navigation affect all of these outcome measures and are important in understanding consumer behavior in commercial Web sites. The research has marketing significance for both academics and industry practitioners interested in the commercialization of the WWW. Knowledge of the relationship between surfersí inclinations and reactions may lead to better Web site design.

3. Experiential and Goal-Directed Behavior in Computer-Mediated Environments
 Hoffman and Novak (1996) suggest two categories of consumer behavior during time spent in Computer-Mediated Environments: goal-directed and experiential behavior. Goal-directed and experiential behavior are distinguished by "(1) extrinsic versus intrinsic motivation, (2) instrumental versus ritualized orientation, (3) situational versus enduring involvement, (4) utilitarian versus hedonic benefits, (5) directed versus nondirected search, and (6) goal-directed versus navigational choice" (Hoffman and Novak, 1996, p.62).
 Ritualized orientations are a "less intentional and nonselective orientation" (Rubin and Perse 1987, p. 59). In contrast, instrumental orientations are "more intentional and selective," which reflects "purposive exposure to specific content" (p.59). Bloch, Sherrell, and Ridgway (1986) suggest that an ongoing search is a function of enduring involvement with the product (or with the CME), whereas a pre- purchase search is a function of situational involvement with the purchase. Hoffman and Novak (1996) describe two behavioral categories of Web users:

 The corporate buyer using the Web to close a deal for computer components experiences an extrinsically motivated, instrumental, goal-directed flow state. On the other hand, net surfers exploring the Web in their daily quest for the latest interesting sites experience an intrinsically motivated, ritualized, experiential flow state. It is important to recognize that consumers engage in both goal-directed and experiential behaviors, flow may occur with both types of behaviors, and the optimal design of a CME site differs according to whether the behavior is goal-directed or experiential.

 Goal-directed surfers are characterized by situational involvement and aim to complete a  specific task. However, consumers who navigate a CME exhibit enduring involvement with an interest area or computers for their hedonic benefits. They search nondirectly and recreationally (Bloch, Sherrell, and Ridgway 1986; Csikszentmihalyi 1983).
 Although surfersí experiential behavior is a very important category to research consumer behavior in a CME, only goal-directed behavior of surfers is considered in this study for two reasons. First, surfersí attribution according to their behavior pattern will be totally different. It is assumed that, compared to goal-directed surfers, surfers who do nondirected searches with navigational choice have different attribution for their Web searching performance. Surfersí experiential behavior will be excluded in this study, because it not only has its own important research field but also requests different experimental processing. Second, the main reason why surfers search special commercial Web sites is to get information about products for which they are looking. Also, it is assumed that goal-directed searching has more direct relation to purchase intention than experiential searching. Experiential behavior with enduring involvement will be considered as the research question to be solved in the next study.

 4. Theoretical Foundations
 4.1 Understanding Attribution Theory
 Attribution Theory
 Attribution theory relates to the processes individuals use to interpret events. That is, it explains how people develop ideas of causation concerning the outcomes of their own behavior and that of others. According to Weiner (1972), attribution theorists deal with "why" questions as opposed to "what if" or other prediction- oriented questions. They are engaged in the relationship between phenomena (e.g., behavior, effects, events) and the causes of those phenomena (Teas and McElroy, 1986,p.76).
 Kelley (1973,p.127) argues that causal attributions play a vital role in providing the inputs to and the basis for deciding the course of action (Teas and McElroy, 1986,p.76). Attribution theory is based upon three basic assumptions: First, individuals will attempt to assign causes for important instances of behavior and, when necessary, seek additional information in order to do so. Second, individuals will assign causal explanations in a systematic manner. Finally, the particular cause that an individual attributes to a given event has important consequences for his or her subsequent behavior (Jones et al. 1972.p.x).
 Weiner (1985,1986) argues that three dimensions of causality have been identified with some degree of regularity: (a) locus, (b) stability, and (c) controllability.
 The locus dimension of causality, derived from the study of Heider (1958), assumes that a performance can be attributed either to internal individual factors or external environmental factors. External attributions would include luck or task difficulty; internal factors are effort, mood, ability, etc.
 The stability of causality, developed by Weiner (Weiner et al. 1972), refers to whether the cause is perceived to be relatively enduring (stable) or changing (unstable) from one situation to another. That is, distinctions are made between relatively permanent factors such as ability, task difficulty and personality, and relatively temporary factors such as effort, luck, and mood (Weiner 1979, 1980).
 Controllability of causality refers to the degree of volition that can be exerted over a cause (Onifade, Harrison, and Cafferty, 1997, p442). Causes may be volitional (choice can be involved) or nonvolitional (constraints may force a product failure). For example, suppose a consumer fails to find useful information in a commercial Web site. This could be due to a controllable cause - a Web surfer making no effort to find it properly - or to an uncontrollable cause- the Web surfer is blind. Similarly, when a Web userís search fails because the computer hardware has a problem, the cause is under the control of an external variable (computer); when the poor search is due to a fire in the computer lab, the cause is not under the control of the computer.
 Locus and stability are the most popularly cited and most heavily researched causal dimensions. They will comprise the major frame of this study because uncontrollable causes such as the blind Web surfer or the fire in the computer lab are beyond the limits of the research question.
 Self-Serving Bias of Attribution
 Surfersí high performance on Web searching will be perceived as more internal than their low performance. Is it true ? Several attributional researchers (Bettman and Weitz, 1983; Bradley, 1978; Clapham and Schwenk, 1991; Miller and Ross, 1975; Zuckerman, 1979) have recommended ëself-serving biasí as the answers.
 The attributional explanation suggests that people take credit for high performance and blame external factors for low performance to protect their self-esteem. Thus, the self-serving bias reflects the tendency of individuals to attribute high performance to factors associated with the individual (ability, effort) and low performance to external factors (task difficulty, luck). The effects of this bias are evident in peopleís explanations for their success/failure, as well as the success/failure of unknown individuals (Kelley and Michela, 1980). Organizational researchers have reported self-serving attributions in managersí explanations of firm performance (Bettman and Weitz, 1983; Clapham and Schwenk, 1991).

 Expectancy Beliefs and Performance-Attribution-Expectancy Linkages
 How would commercial Web surferís performance and their attribution of a performance affect subsequent expectancy estimates such as their continuing consideration of Web sites as a commercial inquiry tool ? The previous studies of a personís expectancy beliefs provide some clues.
 Because a personís expectancy beliefs can be predicted to be positively related to his/her level of motivation, the examination of factors influencing expectancy beliefs is useful. Expectancy motivation models have been heavily researched as the empirical framework of sales management or consumer behavior. Expectancy motivation studies have served as a predictor of human behavior about the effort and performance across occupational groups (Campbell et al. 1970; Lawler  1968).
 Walker, Churchill, and Ford (1977) defined the concept of expectancy as follows:
...the salesmanís estimate of the probability that expending a given amount of effort on task (I) will lead to an improved level of performance on some performance dimension (j).

 In the strict sense of the conceptional distinction, Teas and McElroy (1986, p.78) differentiate expectancies from expectations.
Expectations are related to aspirations and reflect anticipated outcome levels (e.g., expected performance levels). Expectancies, on the other hand, refer to the effort-performance linkage. Thus, expectancy refers to the perceived relationship between effort and some specified or anticipated level of performance.

 As shown in Table 1, Teas and McElroy (1986, p.78) also developed Performance-Attribution-Expectancy linkages to explain the logical connection among consumerís performance, attribution and expectancy on their future behavior. The basic assumption of performance-attribution-expectancy linkages is that performance level leads to an attribution process, which, in turn, affects expectancy on the future task. Thus, the effect of performance on expectancy beliefs is presumed to be subject to the level of performance and the causal attribution used to rationalize a given level of performance.
                                      ------
      Table1
      --------

 Under conditions of successful performance, expectancy estimates are hypothesized to increase positively if the high performance is attributed to (a) external-stable attribution, (b) internal-stable attribution, or (c) internal-unstable attribution. Teas and McElroy (1986, p.77) provide one example:
Salespersons can be expected to attach high expectancy estimates to future task performance in those where they take credit for successful past performance on that task (make an internal attribution to ability or effort) or where they perceive the task as easy (make an external-stable attribution).  However, if the successful performance is attributed to external-unstable causes (luck), no evidence is provided concerning any systematic effort-performance linkage.

  After poor performance or failure, positive expectancy estimates are hypothesized to increase only if the low performance is attributed to internal-unstable causes (i.e., a lack of effort). However, if low performance is attributed to stable factors such as low ability (internal-stable causes) or high task difficulty (external-stable causes), the linkage between effort and future tasks can be expected to be negative. Finally, if low performance is attributed to an external-unstable attribution (luck), poor performance will provide no expectancy estimate. Consequently, poor performance attributed to external-unstable factors can be expected to have a negligible effort-performance linkage.

 4.2 Extracting Attribution Variables in Computer-Mediated Environments
 In this section, surfersí conceptual factors of network navigation in a CME (Hoffman and Novakís conceptual model, 1996; Novak, Hoffman and Yungís causal model, 1997) will be collected and assigned as two attribution dimensions; locus and stability (Weiner 1985, 1986; Teas and McElroy, 1986). These selected variables in a CME will be re-measured as surfersí attribution of their performance. Two studies will be used to verify the extracted variables and to find the Web surfersí attribution. The first study will collect surfersí attribution for why they failed or succeeded in finding the commercial information they were looking for. Using factor analysis, the second study will group these attribution items to extract surfersí attribution factors. These derived attribution factors will be matched with the selected variables in network navigation of a CME.

 Hoffman and Novakís (1996) Conceptual Flow Model of Flow of Network Navigation in CMEs

 The definitions of "flow" have been proposed by Trevino and Webster (1992) and Csikszentmihalyi & Csikszentmihalyi (1988). Trevino and Webster (1992) operationally define flow as the linear combination of four characteristics: control, attention, curiosity, and intrinsic interest. Csikszentmihalyi & Csikszentmihalyi (1988) focused on the congruence of a personís skills in a given activity and perceptions of the activityís challenges. Hoffman and Novak (1996)  believe the concept of flow in Computer-Mediated Environments underlies many crucial components of the consumerís interaction with the firm and its offerings.
 Hoffman and Novak developed the flow model of network navigation in Computer-Mediated Environments. They explained that flow consists of a process state requiring a set of antecedents to occur and results in a set of consequences. They mapped the issues on the boxes and routes of Figure 1(see Appendix) to an outline of how they relate to the process model of network navigation in a CME. They explain their model in the following way (Hoffman and Novak, 1996, p.58):
For expository purposes, we diagram only the most important links. The model shows neither the complex set of feedback loops and pathways nor the fully dynamic nature of the process. A consumer enters the hypermedia CME and engages in network navigation. There are several points of exit from the environment, as well as opportunities to continue navigation, with flow in essence serving as the glue holding the consumer in the hypermedia CME.

 They also suggest research issues such as primary antecedents, secondary antecedents of flow, and surfersí heterogeneity.
 1. Flow
 Ghani, Supnick and Rooney (1991) define two key characteristics of flow as "the total concentration in an activity and the enjoyment one derives from an activity...the precondition for flow is a balance between the challenges perceived in a given situation and skills a person brings to it (p.230)" Trevino and Webster (1992) explain flow in the following way:
"flow represents 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 userís curiosity is aroused during the interaction, and (d) the user finds the interaction intrinsically interesting..(p.542)"

 Hoffman and Novak (1996) define surfersí 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 (Hoffman and Novak, 1996, p.57)."
 2. Primary antecedents of flow
 Two primary antecedents must be present in adequately motivated Web users of Computer-Mediated Environments for the flow experience to occur: (1) skills and challenges must be perceived as congruent and above a critical threshold and (2) focused attention must be present (Hoffman and Novak, 1996).
 (a) Skills and challenges - Hoffman and Novak (1996) define skills as the consumerís capacities for action and challenges as the opportunities for action available to the consumer in a CME. They also explain the relationship between skills and challenges in a CME (Hoffman and Novak, 1996, p.60)
Only when consumers perceive that the hypermedia CME contains challenges congruent with their own skills can flow potentially occur. If network navigation in a CME  does not provide for congruence of skills and challenges, then consumers either become bored (i.e., their skills exceed the challenges) or anxious (i.e., the challenges exceed their skills) and either exit the CME or select a more(or less) challenging activity within it.

 (b) Focused attention - focused attention is also necessary to experience flow. Focused attention is defined as "a centering of attention on a limited stimulus field" (Csikszentmihalyi 1977, p.40). In Figure 1,  Hoffman and Novak(1996) indicate the content characteristics of interactivity and vividness in attracting attention. "The performance characteristics of ease of access, mapping, speed, and range all combine to increase interactivity (Hoffman and Novak, 1996,p.61)." Vividness is characterized as "the representational richness of a mediated environment as defined by its formal features" (Steuer 1992, p.81), such as its breadth and depth. Breadth is the number of sensory dimensions presented and influence media concurrency (Valacich et al. 1993) and media richness (Daft and Lengel 1986; Daft, Lengel, and Trevino 1987). Depth is the resolution or quality of the presentation (Steuer 1992) and is highly related to media bandwidth. Hoffman and Novak (1996, p.61) explain the relation between skills, challenges, and attention in a CME:
 Consumers must focus their attention on the interaction, narrowing their focus of awareness so that irrelevant perceptions and thoughts are filtered out, and they must perceive a balance between their skills and the challenges of the interaction.

 3. Secondary antecedents of flow
 Two additional antecedents: interactivity and telepresence increase the subjective intensity of the consumerís flow state. However, they are not sufficient alone to produce a flow state. A strong sense of telepresence is induced by vividness and interactivity (Sheridan 1992), as well as focused attention. That is, telepresence is a complex concept that originated not only from a userís internal variable (focused intention) but also external characteristics of a CME such as vividness and interactivity.

 4. Surfersí heterogeneity
 Hoffman and Novak (1996) defines the autotelic personality and surfersí optimal stimulation level as surfersí heterogeneity. Web users are heterogeneous in their ability to experience flow in a CME. A Web surfer who has the autotelic personality is a person "who is able to enjoy what he is doing regardless of whether he will get external rewards from it" (Csikszentmihalyi, 1977, p.22) and "who thus is more likely to experience flow for a given activity (Hoffman and Novak, 1996, p.61)."
 There are individual differences in the Optimal Stimulation Level. People having a higher OSL present exploratory behavior (Raju 1980) and increased curiosity-motivated, variety-seeking, risk-taking behavior (Steenkamp and Baumgartner 1992). Hoffman and Novak (1996) expect that people with higher OSLs are more likely to possess the autotelic personality trait and experience flow in a CME.

 Research Issues in a Conceptual Model of Flow and Surfersí Attribution in CMEs

 As we examined above, two dimensions of causality: locus, and stability have been identified in attribution theory. The locus dimension refers to whether the cause of an event is perceived to be internal or external to the person. The stability of causality refers to whether the cause is perceived to be relatively enduring (stable) or changing from one situation to another (unstable). As see in Table 1, ability would be an example of an internal stable factor, and effort would be an example of an internal unstable factor. The typical examples of external stable and unstable factors are task difficulty and luck, respectively.
 As see in Table 2, characteristics in a process model of network navigation can be allocated by two causal dimensions (locus and stability). Surfersí internal stable factors include surfersí heterogeneity (autotelic personality and optimal stimulation level) and control characteristics (skills and challenges). On the other hand, content characteristics (interactivity, vividness) and performance characteristics (speed, mapping, range, information accessibility) will be attributed to external variables. However, surfersí flow and focused attention will be presented differently in the diverse Web searching performance. Therefore, we can attribute flow and focused attention to surfersí internal unstable factor.
 Finally, telepresence is a complex concept of an internal variable (focused attention) and external variables (vividness, interactivity).
     ---------
     Table 2
     ---------

 Research Issues in a Causal Model of Flow and Surfersí Attribution in CMEs
 Based upon Hoffman and Novakís (1996) conceptual model of flow, Novak, Hoffman and Yung (1997) tried to develop a causal model of flow in CMEs. After testing several models, they derived a causal model (Figure 2, see Appendix). In Figure 2, Novak, Hoffman and Yung (1997) include control, which refers to Azjenís (1988) construct of perceived behavioral control. Control is indicated as a consequence, rather than antecedent, of flow.
 Two additional variables are included as predictors of skill: when the respondent first reported using the Web and the amount of time per day the respondent reported using the Web. Their revised model also suggested interactivity is best thought of as three separate components (speed, range, and mapping), which are somewhat correlated but affect different constructs (Novak, Hoffman and Yung, 1997). That is, speed affects challenge, play, and focused attention. Range (the number of actions available at a given time) affects challenge. Mapping (the naturalness of the interaction) affects play. In addition, they include several latent variables such as play, arousal, and time distortion as antecedents of flow.
     ---------
     Table 3
     ---------

 Latent variables in a causal model of network navigation (Table 3) are allocated by two causal dimensions (locus and stability). Interactivity (speed, mapping, range) can be attributed to external variables. Surfersí flow, time distortion, arousal, focused attention, and control are presented diversely in the various performance of the Web searching. Therefore, these are all internal unstable factors. On the other hand, surfersí skill, challenge, play and optimum stimulation level, , are internal stable variables because they originally come from surfersí personality.

 Research Issues in Flow Channel Segmentation Models
 Novak and Hoffman (1997) suggest that flow channel segmentation models are based upon Csikszentmihalyiís definition of flow in terms of skills and challenges. However, the segmentation models attempt to account for all possible combinations (channels) of high/low skills and challenges.
 After several empirical tests about flow channel models (e.g., Ellis, Voelkl & Morris 1994; and LeFevre 1988), the eight channel model (Massimini & Carli 1988; Ellis, Voelkl & Morris 1994) was developed. It allows for intermediate (moderate) levels of skills and challenges and identifies four additional channels: arousal, control, relaxation, and worry. Flow is defined as high skills and high challenges; apathy as low skills and low challenges; anxiety as high challenges and low skills; and boredom as high skills and low challenges. The arousal vs. relaxation distinction is collinear with challenge, and the control vs. worry distinction is collinear with skill.
 Massimini and Carli (1988) examined mean ratings on 23 scales by the eight segments with a sample of 47 Italian students aged 16 to 19. Their data support the eight channel model. That is, Novak and Hoffman (1997) suggest that their results show a clear differentiation between items that point toward control and arousal. But, their principal components analysis of the means of the 23 items on the eight flow channels shows that  23 scales explain only six flow channels, such as control vs. worry, flow vs. apathy and arousal vs. relaxation. The identical factors with boredom vs. anxiety are not represented. Therefore, the boredom vs. anxiety, an overlapping concept with six flow channels, is excluded in this study.

 Surfersí Conceptual Variables and Attribution in CMEs extracted from the previous research

 Based upon Hoffman and Novakís conceptual model (1996) and Novak, Hoffman, and Yungís causal model (1997), surfersí conceptual factors is derived. Latent variables in a process of network navigation are redistributed into two causal dimensions (locus and stability) in order to understand surfersí attribution toward their Web searching performance. That is, all latent variables considered for this study will be allocated by four dimensions­ internal stable, internal unstable, external stable and external unstable­ and verified by Web surfers. Based upon this reallocation of latent variables in a process of network navigation, research hypotheses that are driven from attribution theory will be tested.
 As Table 4 indicates, latent factors of flow state in a CME are derived from Hoffman and Novakís conceptual model (1996), and Novak, Hoffman, and Yungís causal model (1997).
     --------
     Table 4
     ----------

 For this study, skills and challenges are selected as surfersí internal stable factors.
Hoffman and Novak (1996) define skills as the consumersí capacities for action. Also, they explain challenges as the opportunities for action available to the consumer in a CME.
However, Novak and Hoffman (1997) define challenges as the internal stable factor that originally existed in surfersí attitudes in their flow channel segmentation models, based upon Csikzentmihalyiís definition of flow in terms of skills and challenges.
 Hoffman and Novak (1996) define the autotelic personality, and surfersí optimal stimulation level as surfersí heterogeneity. Web users are heterogeneous in their ability to experience flow in a CME. A Web surfer who has the autotelic personality is a person "who is able to enjoy what he is doing regardless of whether he will get external rewards from it" (Csikszentmihalyi, 1977, p.22). People having a higher OSL present exploratory behavior (Raju 1980) and increased curiosity-motivated, variety-seeking, risk-taking behavior (Steenkamp and Baumgartner 1992). Hoffman and Novak (1996) expect that people with higher OSLs are more likely to possess the autotelic personality trait and experience flow in a CME. Surfersí heterogeneity, such as the autotelic personality and OSL, is excluded from the research issue for this study because these are "the antecedents of primary antecedents of flow," such as skills and challenges. Therefore, skills and challenges are sufficient factors to examine surfersí internal stable variables. Also, examining too many variables would deviate from the main purpose of the research.
 Focused attention, control, arousal, and flow are selected as the internal unstable factors. Focused attention is defined as "a centering of attention on a limited stimulus field" (Csikszentmihalyi 1977, p.40). Control is the opposite concept of worry and happens when skills are high and challenges are low. At first, Novak, Hoffman, and Yung (1997) define control from Azjenís (1988) construct of perceived behavioral control and classify it as a consequence, rather than antecedent, of flow. But they, later, suggest that control is either an antecedent or consequence of flow because the role of perceived behavioral control in Hoffman and Novakís (1996) conceptual model conflicts with the role of control as an antecedent of flow in the other research. Control will be considered as an antecedent of flow and surfersí internal unstable factor in this study because the surfersí controllability of Web-searching is not stable. Arousal is the opposite concept of relaxation and happens under the circumstance of high challenges and low skills. Arousal is also an internal unstable factor for this study. Flow is defined 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" (Hoffman and Novak, 1996). Flow is an internal unstable factor because the level of flow is changeable whenever surfers are searching. Novak, Hoffman, and Yung (1997) suggest that play, while initially hypothesized to function identically to flow, can also be an immediate antecedent of flow. Play, however, is excluded from the research because this study will examine surfersí goal-directed behavior.
 Interactivity (speed, mapping, range) and information accessibility are selected as the external stable/unstable variables for a causal test model in CMEs by surfersí attribution. Accumulated industry experience and anecdotal evidence strongly support the contention that the primary barrier to consumer adoption of the Web as a commercial medium is ease of access. Convenience of access is at the core of the adoption of any technological application and determines its ultimate success (Gupta, 1995). Information accessibility is specifically selected as an important external factor because this study focuses on surfersí goal-directed behavior. Hoffman and Novak (1996) indicate that the performance characteristics of ease of access, mapping, speed, and range all combine to increase interactivity. Range is the number of actions available, and mapping means the naturalness of the interaction. Interactivity and information accessibility may be divided into two branches: interactivity and information accessibility at a general CME and interactivity and information accessibility at a given time during the Activity/Survey. The former factors can be considered as the external stable factors while the latter variables are the external unstable variables. Therefore, information accessibility and interactivities such as range, mapping, and speed at a general CME are selected as the external stable variables, while information accessibility and interactivities such as range, mapping, and speed at a given time during the Activity/Survey are the external unstable variables for this study.
 A strong sense of telepresence is induced by vividness and interactivity (Sheridan 1992), as well as focused attention. That is, telepresence is a complex concept that originated not only from a userís internal variable (focused intention) but also external characteristics of a CME, such as vividness and interactivity. Therefore, telepresence is excluded in this study.
 Vividness is defined as "the representational richness of a mediated environment as defined by its formal features" (Steuer 1992, p.81), such as its breadth and depth. Vividness is ruled out in this study because it is an overlapping concept with mapping, which is the naturalness of interactivity. Time distortion is also excluded in order to avoid the duplication of the concept with focused attention.

 4.3 Verification of the Extracted Attribution Variables through Factor Analysis

 Two studies were executed to verify the extracted variables using factor analysis. Following the findings of previous researchers (Hoffman and Novakís conceptual model, 1996;  Novak, Hoffman and Yungís causal model, 1997) on latent factors of flow state in a CME and two attribution dimension of locus and stability (Weiner 1985, 1986 and Teas and McElroy, 1986), it was expected that skill and challenge would be the surfersí internal stable factors; followed by focused attention, control, arousal and flow as the internal unstable factors; an interactivity (speed, mapping, range) and information accessibility as the external stable/unstable variables. However, even if the previous factors of flow state in a CME contained all conceptual variables occurring from surfersí searching, they are not the Web usersí  attributing variables that influence their searching performance. Moreover, as said in the beginning of the paper, only goal-directed behavior of surfers is considered for this research. Therefore, these selected variables in a CME have to be re-measured and verified as surfersí attribution factors toward their performance.

 STUDY 1
 Using mall intercepts in Columbia, Missouri, the first study was  done in May 1998 to ask the main reasons why Web users failed or succeeded in finding the commercial information for which they were looking. 104 respondents were asked to answer an open-ended questionnaire. The sample was evenly divided between men and women. Several researches (Graphic, Visualization, and Usability Center, 1995; Gupta, 1995; Vonder Haar, 1995; Zeigler, 1995) commented on WWW surfers:
 
The PC market is "young," since 58 percent of PC owners have had their PCs for less than two years (Zeigler, 1995). According to the Odyssey Homefront Survey Results (Vonder Haar, 1995), one-third of U.S. households have a PC at home, and 18% have modems. Web surfersí average age (Graphic, Visualization, and Usability Center, 1995; Gupta, 1995) is 34. 91% among them have some college education or better, the median income is between $ 50,000 and $ 60,000.
 SRIís (1995) analysis of the Web population provides psychographic insight of the Web surfer. It identifies two broad categories of the Web surfer. The first group called the "upstream" audience is 50% of the current Web population. This group, which represents 10 % of the U.S. population, is successful men and women with high self-esteem (they are  "Actualizers" by SRI terms). According to SRI, the second group will determine the possibility of the Web as a commercial medium. SRI refers to them as strivers and experiencers. Experiencers are impulsive and enthusiastic, seeking excitement from their life, while strivers seek approval from the world around them.

 Surfersí ages and Web experience in the sample would reasonably represent the adult user base of the Web at the time of the study. Eighty-five percent of respondents were experienced users of commercial Web searching. Seventy-four respondents (71%) said that they use the Internet more than 30 minutes a day. The average age was 30.
 As we can see in Table 5, three doctoral students majoring in journalism, computer engineering and social-psychology specified 51 items from the respondentsí responses. The ordering of items followed their frequency. Many items were similar in terms of their conception, hence factor analysis, the second study, will group the duplicate meanings from items.
 Nineteen items overlapped with the variables derived from the previous research. Interestingly, except only 2, 12, 46 and 51 items (see table 5), respondents did not separate commercial Web searching from their general Web searching. That is, researchers could not find the unique items only from commercial Web searching, meaning that the general variables in CMEs, such as technical traits and surfersí skill, affect not only surfersí experiential behavior but also their goal-directed behavior.
      ---------
      Table 5
      ---------
 

 STUDY 2
 As mentioned earlier, the selected variables from flow state in a CME have to be re-measured and verified as surfersí attribution factors toward their performance. Fifty-one items from the first pilot study were collected. The second study was executed to group these duplicated items.
 A group of college students is an appropriate sample to analyze Web usersí attitudes. Using personal interviews and e-mail surveys at the University of Missouri-Columbia, a study was done in September 1998. One hundred-fifty participants completed the questionnaire in personal interviews. Of these, 29 were omitted from the analysis as a result of inconsistencies or incomplete questionnaires. One hundred-eight e-mail users responded to the questionnaire. Of these, 16 also were omitted from the final analysis. A total of 213 participants (121 interviewees and 92 e-mail respondents) were selected, and their responses were analyzed to group Web usersí attribution factors toward their performance. Of the selected respondents, 92% were experienced users of the Web and 71% had searched commercial Web sites for the purpose of purchasing products or collecting information.
 Factor analysis was used to examine the groupings among the rating scales. Table 6  shows the factor analysis results for a varimax rotation. Following the "break-in-the-roots" method described by Gorsuch (1974), it was determined that factors showing contributions to the total variance of 5% or larger would be included in the subsequent analysis. Six factors were near or above the cut-off 5 percent contributions to the total variance. Also, those factors were analyzed with a 1.0 cut-off for the eigenvalues. Once identified, these conceptual areas were compared with the factors of flow state in a CME identified in previous research.
 As shown in Table 6, the first factor accounted for 23.7 %  of the variance. The  leading items included "I am very skilled/ not very skilled at using the Web," "I am experienced/ not experienced," "I have precise search keywords/ poor choice of search words." The emergence of a first factor involving a userís skill is consistent with previous studies such as the 7th GVU Survey and Novak, Hoffman and Yung, 1997. Thus, skills (Hoffman and Novak, 1996) as the consumersí capacities for action were defined not only as a conception in flow state in CMEs but also a factor that Web surfers attribute their high or low performance.
 Factor 2 from the various computer capability centers on its mapping. The leading items included not surfersí characteristics but computer characteristics such as the linking of computer, readability of Web browsers and Web design. Novak, Hoffman and Yung (1997) specified the general characteristics of a computer as interactivities; mapping, range and speed. Factor 2 is mainly consistent with mapping, but it also contains range items. That is, mapping and range as interactivities were not separated from the second study. Factor 2 will be identified as "mapping," comprising its leading items as mapping distinctions.
 Factor 3 calls attention to the usersí challenge in their Web searching. The items in this factor relate to the dimension called challenge used in 7th GVU Survey. Novak and Hoffman (1997) defined challenges as the internal stable factor that originally existed in surfersí attitudes. Factor 3 also will be named as surfersí attribution toward their challenge.
 The leading items of Factor 4 were modem velocity, and the working speed of the computer server. We call it overall computer speed. Also, it is consistent with speed among computer interactivities used in Novak, Hoffman, and Yung (1997).
 Flow (Hoffman and Novak, 1996) is defined as surfersí unconscious state occurring during their network navigation, characterized by a seamless sequence of responses facilitated by machine interactivity. It also accompanies intrinsically enjoyable feeling. As shown in Table 6, Factor 5, involving enjoyable and unconscious immersion to the fast pace of the computer, will be designated as "flow," because it is consistent with three items of apathy vs. flow used in Novak, Hoffman and Yung (1997).
 Factor 6 items relate to three items of focused attention based upon Ghani and Deshpande (1994), and modified from Novak, Hoffman, and Yung (1997). Focused attention is a centering of attention on a limited stimulus field (Csikszentmilhalyi, 1977). Thus, Factor 6 is defined as a surfersí attitude that attributes their high or low performance to their focused attention.
 Whereas control is the opposite concept of worry and happens when skills are high and challenges are low, arousal is the opposite concept of relaxation and happens under the circumstance of high challenges and low skills. Relaxation vs. arousal, worry vs. control and information accessibility derived from items used in Novak, Hoffman, and Yung (1997) did not emerge as factors because their concepts overlapped with other factors such as challenge, skill and interactivity. For example, whereas one item from worry vs. control; "I get/ donít get frustrated while I am searching the Web" is designated as factor 1; surferís attribution toward their skill, the other; "The information that one can access with the Web is unlimited/limited" is included in factor 2; surferís attribution toward computer mapping.
 The second study tried to re-measure and verify the previous factors of flow state in a CME, which contained all conceptual variables occurring from surfersí searching, as surfersí attribution factors toward their performance. As explained above, six factors are derived. Following the emergence of six attribution factors and two attribution dimensions; locus and stability (Weiner 1985, 1986 and Teas and McElroy, 1986), it is expected that skill and challenge would be the surfersí internal stable factors; followed by focused attention and flow as the internal unstable factors; and interactivity (speed, mapping) as the external stable/unstable variables.
 Using factor analysis, however, this researcher could not categorize surfersí attribution variables with four sections (internal stable, internal unstable, external stable and external unstable factors) because surfersí attitude, which attributes computer speed and mapping to their performance, would be the external stable/unstable variables. Hence a third study will be necessary to specify surfersí attribution factors with the four categories.
 Through a mall intercept interviews, the third study will retest that skill and challenge would be the surfersí internal stable attribution factors, and focused attention and flow would be the internal unstable attribution variables. Also, after the researcher voluntarily divides computer speed and mapping as two dimensions; computer mapping and speed at a general CME, and computer mapping and speed at a given time during the Activity/Survey, respondents will asked to evaluate its justification. Finally, using the Activity/Survey designed as a laboratory experiment, research hypotheses will be tested based upon attribution variables in a process of network navigation and performance-attribution-expectancy linkages (See Table 7).
     ---------
     Table 6
     ---------
 

 5. Research Hypotheses

 If the attribution process influences surfersí perceptions of event causes, then this ëself-serving biasí should be evident in surfersí explanation of high and low performance on a Web searching tasks. That is, when their performance on a Web searching tasks is high, the locus of the performance should be attributed more to internal than external causes. On the other hand, when their performance is low, the event locus should be seen as more external than internal. This attribution study thus suggests the following hypothesis:

H1: Surfersí high performance on a Web searching tasks will be perceived as more internal than their low performance.

 To test performance-attribution-expectancy linkages of Web surfers, it is suggested here that causal attribution processes mediate the feedback relationship between performance and expectancy estimates. The following hypotheses specify the proposed relationships between performance attributions and Web surfer expectancy estimates:
     ---------
     Table 7
     ---------

 
H2: Stable attributions (internal or external) will result in increased expectancy estimates for the Web searching task following high performance and decreased expectancy estimates following low performance. As indicated in Table 7, this proposition suggests the following:

 H2- aa, ab. Low performance on the Web searching task attributed by  commercial Web surfers to an external-stable cause (interactivities such as mapping and speed at a general CME) will have a negative impact on their expectancy estimates for future Web searching.
 
 H2- ba, bb. Low performance on the Web searching task attributed by commercial Web surfers to internal-stable cause (skills, challenges) will have a negative impact on their expectancy estimates for future Web searching.

 H2- ca, cb. High performance on the Web searching task attributed by commercial Web surfers to external-stable cause (interactivities such as mapping and speed at a general CME) will have a positive impact on their expectancy estimates for future Web searching.

 H2- da, db. High performance on the Web searching task attributed by commercial Web surfers to internal-stable cause (skills, challenges) will have a positive impact on their expectancy estimates for future Web searching.

 
H3: Internal-unstable attributions will result in increased expectancy estimates. As indicated in Table 7, this proposition suggests the following:

 H3- aa, ab. Low performance on the Web searching task attributed by commercial Web surfers to internal-unstable cause (focused attention and flow) will have a positive impact on their expectancy estimates for the future Web searching.

 H3- ba, bb. High performance on the Web searching task attributed by commercial Web surfers to internal-unstable cause (focused attention and flow) will have a positive impact on their expectancy estimates for the future Web searching.

H4- a, b. External-unstable attributions (interactivities such as mapping and speed at a given time during the Activity/Survey) will not be related to expectancy estimates.

 6. Research Methods
 The Activity/Survey method will be used for this study. This method is selected for two main reasons.
 First, the Activity/Survey method is useful for laboratory experiments and either concurrently or retrospectively determining the experience of flow for specific events (Novak, and Hoffman, 1997). The problem is whether respondents can reliably evaluate flow after, rather than during, an activity. However, Novak, and Hoffman (1997) suggest that surveys completed immediately after completion of an activity have greater validity than surveys describing an activity in which the person has engaged for a long period of time.
 Second, it is natural that surveys after completion of an activity are executed because surfersí attributions toward their performance level have to be examined.
 
 

     References

Alwitt, Linda F., and Paul R. Prabhaker (1994), "Identifying Who Dislikes Television Advertising: Not By Demographics Alone," Journal of Advertising Research , 34, 6, 17-29.

Ajzen, Icek (1988), Attitudes, Personality, and Behavior. Chicago, IL: Dorsey Press.

Bare, S. L. (1996), Capabilities of Commercial Sites on the World-Wide Web: An Exploration of Depth of Adoption. Unpublished masterís thesis, University of Alabama, Tuscaloosa.

Bettman, J. R. and B. A. Weitz (1983), "Attributions in the board room: causal reasoning in corporate annual reports," Administrative Science Quarterly, 28, 165-83.

Berthon, P., Leyland F. Pitt, and Richard T. Watson (1996), "The World Wide Web as an Advertising Medium," Journal of Advertising Research , 36, 1, 43-54.

Becker, L. B., R. A. Martino, and W. M. Towners (1976), "Media Advertising Credibility," Journalism Quarterly, 53, 216-22.

Bloch, Peter H., Daniel L. Sherrell, and Nancy M. Ridgway (1986), "Consumer Search: An Extended Framework," Journal of Consumer Research, 13 (June), 119-26.

Bradley, G.W. (1978), "Self-serving biases in the attribution process: a reexamination of the fact or fiction question," Journal of Personality and Social Psychology, 36, 56-71.

Campbell, J. P.., M. D. Dunnette, E. E. Lawler, and K. E. Weick, Jr. (1970), Managerial Behavior, Performance, and Effectiveness, New York: McGraw-Hill.

Clapham, S. E. and C. R.Schwenk (1991), "Self-serving attributions, managerial cognition, and company performance," Strategic Management Journal, 12, 219-29

Csikszentmihalyi, Mihaly (1977), Beyond Boredom and Anxiety. San Francisco: Jossey-Bass.

Csikszentmihalyi, Mihaly (1983), "Measuring Intrinsic Motivation in Everyday Life," Leisure Studies, 2 (May), 155-168.

Csikszentmihalyi, Mihaly and Isabella Selega Csikszentmihalyi (1988), Optimal Experience: Psychological Studies of Flow in Consciousness. Cambridge: Cambridge University Press.

Daft, Richard L. and R. H. Lengel (1986), "Organizational Information Requirements, Media Richness and Structural Design," Management Science, 32(5), 554-71.

Eighmey, J. (1997), "Profiling User Responses to Commercial Web Sites," Journal of Advertising Research, (May, June), 59-66.

Ellis, Gary D., Judith E. Voelkl, and Catherine Morris (1994), "Measurement and Analysis Issues with Explanation of Variance in Daily Experience Using the Flow Model," Journal of Leisure Research, 26 (4), 381-391.

Ghani, Jawaid A. and Satish P. Deshpande (1994), "Task Characteristics and the Experience of Optimal Flow in Human-Computer Interaction," The Journal of Psychology, 128(4), 382-391.

Ghani, Roberta Supnick, and Pamela Rooney (1991), " The Experience of Flow in Computer-Mediated and in Face-to-Face Groups," in Proceedings of the Twelfth Internatioal Conference on Information Systems, Janice I. DeGross, Izak Benbasat, Gerardine DeSanctis, and Cynthia Mathis Beath, eds. New York: The Society for Information Management, 229-37.

Gooding, Richard Z. and Angelo J. Kinicki (1995), "Interpreting Event Causes: The Complementary Role of Categorization and Attribution Processes," Journal of Management Studies  32:1 January 0022-2380.

Gorn, Gerald J. and Charles B. Weinberg (1984), "The Impact of Comparative Advertising on Perception and Attitude: Some Positive Findings," Journal of Consumer Research, 11(September), 719-727.

Gorsuch, Richard L. (1974), Factor Analysis, W. B. Saunders Company, Philadelphia, PA.

Graphic, Visualization, & Usability Center (1995), GVU Third WWW Usersí Survey
URL: http://www.cc.gatech.edu/gvu/user surveys/User survey Home.html

Gupta, S. (1995), HERMES: A research project on the commercial uses of the World Wide Web. URL: http://www.umich.edu/~sgupta/hermes/

GVU (1997), "GVUís 7th WWW User Survey," http://www.gvu.gatech.edu/user_surveys/survey-1997-04/.

Heider, F. (1958), The Psychology of Interpersonal Relations, New York: Wiley.

Hoffman, Donna L., Thomas P. Novak, and Patrali Chatterjee (1995), "Commercial Scenarios for the Web: Opportunities and Challenges," Journal of Computer-Mediated Communications, Special Issue on Electronic Commerce, 1(3), http://www.usc.edu/dept/annenberg/vol1/issue3/

Hoffman, Donna L. and Thomas P. Novak (1996), "Marketing in Hypermedia Computer-Mediated Environments: Conceptual Foundations," Journal of Marketing, 60 (July), 50-68.

Hoffman, Donna L. and Thomas P. Novak (1996), "Measuring the Flow Experience Among Web Users," http://www2000.ogsm.vanderbilt.edu/novak/flow.july.1997/flow.htm

Hoffman, Donna L. and Thomas P. Novak (1996), "A New Marketing Paradigm for Electronic Commerce," http://www2000.ogsm.vanderbilt.edu/novak/new.marketing.paradigm.html

Internet Advertising Bureau (1998), Internet Advertising Bureau announces first quarter 1998 advertising revenue reporting program results [On-line], http://www.iab.net/news/content/ad_revenue.html

Jones, E. E., D. E. Kanouse, H. H. Kelly, R. E. Nisbett, S. Valins, and B. Weiner (1972), Attribution Perceiving the Causes of Behavior, Morristown, NJ: General Learning Press.

Katz, E. and P. F. Lazarsfeld (1955), Personal Influence, Glencoe: Free Press

Kelly, H. H. (1955), The psychology of personal constructs. New York: Norton.

Kelly, H. H. (1967), "Attribution Theory in Social Psychology," in Nebraska Symposium on Motivation, David Levine, ed., Lincoln, NE: University of Nebraska Press.

Kelly, H. H. (1972), "Attributions in Social Interaction," in Attribution: Perceiving the Causes of Behavior, E. E. Jones, D. E. Kanouse, H. H. Kelly, R. E. Nisbett, S. Valins, and B. Weiner, eds., Norristown, NJ: General Learning Press.

Kelly, H. H. (1973), "The Process of Causal Attribution," American Psychologist, 28(No. 2), 107-128.

Kelly, H. H. and J. L. Michela (1980), "Attribution theory and research," Annual Review of Psychology, 31, 457-501.

Lasswell, H. D. (1948), "The Structure and Function of Communication in Society," in The Communication of Ideas, Bryson, ed., New York: harper and Brothers.

Lawler, Edward E. (1968), "A Correlational-Causal Analysis of the Relationship between Expectancy Attitudes and Job Performance," Journal of Applied Psychology, 52 (No. 6), 426-435.

LeFevre, Judith (1988), "Flow and the Quality of Experience During Work and Leisure." In M. Csikszentmihalyi and I. Csikszentmihalyi (Eds), Optimal Experience: Psychological Studies of Flow in Consciousness, New York: Cambridge University Press.

MacKenzie, S. B., and R. J. Lutz (1989), "An Empirical Examination of the Structural Antecedents of Attitude Toward the Ad in an Advertising Pretesting Context," Journal of Marketing , 53, 2, 48-65.

Massimini, Fausto and Massimo Carli (1988), "The Systematic Assessment of Flow in Daily Experience." In M. Csikszentmihalyi and I. Csikszentmihalyi (Eds), Optimal Experience: Psychological Studies of Flow in Consciousness, (pp. 288-306), New York: Cambridge University Press.

Michalski, J. (1995), "People are the killer app." Forbes ASAP, June 5, 120-122.

Miller, D. T. and M. Ross (1975), "Self-serving biases in the attribution of causality: fact or fiction ?" Psychological Bulletin, 82, 213-25.

Mizerski, R. W.,  L. G. Golden, and Jerome B. Kernan (1979), "The Attributional Process in Consumer Decision Making," Journal of Consumer Research, 2 (September), 123-140.

Novak, T. P., D. L. Hoffman, and Y. F. Yung (1997), "Modeling the Structure of the Flow Experience Among Web Users," Submitted to the INFORMS Marketing Science and the Internet Mini-Conference MIT, March 1998.

Onifade, E., Paul D. Harrison, and Thomas P. Cafferty (1997), "Causal Attributions for Poorly Performing Projects: Their Effect on Project Continuation Decisions," Journal of Applied Social Psychology, 27, 5, 439-452.

Press, Larry (1993), "The Internet and interactive television," Communcations of the ACM, 36 (12), 19-23.

Raju, P.S. (1980), "Optimum Stimulation Level: Its Relationship to Personality, Demographics, and Exploratory Behavior," Journal of Consumer Research, 7 (December), 272-82.

Rubin, Alan M. and Elizabeth M. Perse (1987), "Audience Activity and Television News Gratification," Communication Research, 14 (1), 58-84.

Sheridan, Thomas B. (1992), "Musings on Telepresence and Virtual Presence," Presence: Teleoperators and Virtual Environments, 1 (1), 120-26.

Sparkman, Richard M. and William B. Locander (1980), "Attribution Theory and Advertising Effectiveness," Journal of Consumer Research, 7(December), 219-224.

Steenkamp, Jan-Benedict and Hans Baumgartner (1992), "The Role of Optimum Stimulation Level in Exploratory Consumer Behavior," Journal of Consumer Research, 19 (December), 434-48.

Steuer, Jonathan (1992), "Defining Virtual Reality: Dimensions Determining Telepresence," Journal of Communication, 42 (4), 73-93.

Teas, R. K. and James C. McElroy (1986), "Causal Attributions and Expectancy
Estimates: A Framework for Understanding the Dynamics of Salesforce Motivation," Journal of Marketing, Vol. 50 (January), 75-86.

Tolman, E.C. (1925), "Purpose and cognition: The determinants of animal learning," Psychological Review, 32, 285-297.

Trevino, Linda Klebe and Jane Webster (1992), "Flow in Computer-Mediated Communication," Communication Research, 19 (5), 539-73.

Valacich, Joseph S., David Paranka, Joey F George, and J. F. Nunamaker, (1993), "Communication Concurrency and the New Media: A New Dimension for Media Richness," Communication Research, 20 (2) 249-76.

Verity, J. W., R. D. Hof, E. C. Baig, and J. Carey (1994), "The Internet : How it will change the way you do business. Business Week, November 14, 80-88.

Vonder Haar, S. (1995), "Not so fast : Study on PC market," Inter@ctive Week, September 25, 16.

Walker, O. C., Jr., G. A. Churchill, Jr., and N. M. Ford (1977), "Motivation and Performance in Industrial Selling: Present Knowledge and Needed Research," Journal of Marketing Research, 14(May), 156-168.

Weiner, B. (1972), Theories of Motivation: From Mechanism to Cognition, Chicago: Markham.

Weiner, B. (1979), "A Theory of Motivation for some Classroom Experiences," Journal of Educational Psychology,  71(No. 1), 3-25.

Weiner, B. (1980), Human Motivation, New York: Holt.

Weiner, B. (1985), "An Attributional Theory of Achievement motivations and emotion," Psychological Review, 92, 548-573.

Weiner, B. (1986), An Attributional theory of Motivation and Emotion. New York, NY: Springer-Verlag.

Weiner, B. (1990), "History of motivational research in education," Journal of Educational Psychology,  82, 616-622.

Weiner, B. (1991), "Metaphors in Motivation and Attribution," American Psychologist, Vol. 46, No.9, 921-930.

Ziegler, B. (1995), "Share of homes with Pcs rises to 31% in poll." Wall Street Journal, Feb. 6, Section B, pg. 5, col.1.

Zuckerman, M. (1979), "Attribution of success and failure revisited, or: the motivational bias is alive and well in attribution theory," Journal of Personality, 47, 245-87.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Table 1. Performance-Attribution-Expectancy Linkages
   (R.Kenneth Teas & James C. McElroy, 1986, p.77)

Performance on
Task A
Attribution  Impact on Expectancy Beliefs
            Internal Stable      Positive
               (Ability)
 
          Internal Unstable      Positive
                (Effort)
High
      Performance
           External Stable      Positive
           (Task difficulty)
 
          External Unstable     No impact
                 (Luck)
 
            Internal Stable      Negative
                (Ability)
 
          Internal Unstable      Positive
                (Effort)
Low
         Performance
          External Stable      Negative
        (Task difficulty)
 
         External Unstable     No impact
              (Luck)
 

Table 2. Characteristics of flow state in a CME and surfersí attribution
Characteristics of flow state in a CME  Surfersí attribution
Surfersí heterogeneity Autotelic personality Internal stable
 Optimal stimulation level Internal stable
Control Characteristics Skills Internal stable
 Challenges Internal stable
Content Characteristics Interactivity External variable
 Vividness (breadth, depth) External variable
Performance Characteristics Speed, Mapping, Range External variable
(components to increase interactivity)  Information accessibility
(ease of access) External variable
Process Characteristics Goal-directed or experiential behavior
Flow,  Focused attention Internal unstable
Telepresence A complex concept of internal variable (focused attention) and external variable (vividness, interactivity)
The consequence of flow Consumer learning, Perceived behavioral control, Exploratory behavior, Positive subjective experiences, Distortion in time perception
 

Table 3. Latent variables in a causal model and surferís attribution
Latent variables Surfersí attribution
Antecedents Skill Internal stable
of flow Challenge Internal stable
 Interactivity (speed, mapping,range) External variable
 Telepresence Complex concept
 Arousal Internal unstable
 Focused attention Internal unstable
 Control Internal unstable
 Play Internal stable
 Optimum stimulation level Internal stable
 Time distortion Internal unstable
Flow Internal unstable
Consequences Positive affect Consequence
of flow Exploratory behavior Consequence
 
 
 

Table 4. Latent variables in a conceptual model and causal model and surfersí attribution
Latent variables in a conceptual model Latent variables in a causal model Surfersí attribution
Autotelic personality x Internal stable
Optimal stimulation level Optimal stimulation level Internal stable
Skills Skill Internal stable
Challenges Challenge Internal stable
Interactivity (speed, mapping, range) Interactivity (speed, mapping, range) External variable
Vividness  x External variable
Information accessibility x External variable
Flow Flow Internal unstable
Focused attention Focused attention Internal unstable
Telepresence Telepresence Complex concept
x Play  Internal stable
x Control Internal unstable
x Arousal Internal unstable
x Time distortion Internal unstable
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Table 5. Surferís attribution items of network navigation in CMEs
 Variables Frequency
1. The computer is fast / slow  Interactivity (speed), Novak, Hoffman, and Yung, 1997
2. I am interested/not interested in finding products in the Web sites
3. Most computers today have good/poor search engine
4. The linking of the computer is excellent /terrible
5. There are many relevant /too many irrelevant Web sites through which to search
6. I am very skilled /not very skilled at using the Web - Skill, 7th GVU Survey
7. Using the Web challenges /not challenge me - Challenge, 7th GVU Survey
8. Web searching is not time consuming / too time consuming
9. I find myself lucky /unlucky while I am searching the Web
10. The server of the computer is working well /often down
11. Modem is fast / slow
12. Commercial Web sites today have broad / narrow product ranges
13. I use precise search keywords / I have poor choice of search words
14. I have good / donít have enough instructions on how to search
15. I usually know/ donít know the exact Web address that I have to find
16. I clearly know the right things to do/ I feel confused about what to do while I am searching the Web
17. I am having fun searching the Web sites /  I get tired of Web searching
18. I am calm / I am excited when I am searching Relaxation vs. Arousal, Novak, Hoffman and Yung, 1997
19. Navigating the Web with todayís Web browsers is natural /unnatural Interactivity (Mapping), Novak, Hoffman, and Yung, 1997
20. I know/ donít know my way around on the net
21. I have familiarity / no familiarity with the Internet
22. I am experienced  / not experienced at Web searching
23. I get frustrated / donít get frustrated while I am searching the Web  Worry vs. Control, Novak, Hoffman and Yung, 1997
24. I think Web searching is interesting job / boring job
25. When I am using the Web, I am totally absorbed in /I am not totally absorbed in what I am doing  Focused Attention, Ghani and Deshpande, 1994
26. I am easily distracted / I am not easily distracted while I am searching the Web
27. I am relaxed/ I am stimulated while I am searching the Web   Relaxation vs. Arousal, Novak, Hoffman and Yung, 1997
28. I am impatient / I am patient about it while I am searching the Web
29. Computer downloading takes reasonable time / too much time
30. Web design is very real / Web design is very unreal
31. It is easy /not easy to find a bookmark what I am looking for
32. Readability of Web browsers is powerful / not powerful
33. The number of different ways one can interact with the Web today is limited/ is not limited Interactivity (Range), Novak, Hoffman and Yung, 1997
34. It is easy/ hard for me to find information on the Web Information accessibility, Novak, Hoffman and Yung, 1997
35. I consider/ I donít consider myself knowledgeable about good Web search techniques  Skill, 7th GVU Survey
36. I understand computer language / I donít understand computer language
37. I know / I donítí know the logical steps in Web searching
38. I feel in control of a computer / I feel worried while I am searching the Web  Worry vs. Control, Novak, Hoffman and Yung, 1997
39. Using the Web provides /donít provide a good test of my skills                Challenge, 7th GVU Survey
40. I get much faith /less faith in online transactions
41. I enjoy the fast pace of the computer / I feel apathetic while I am searching the Web  Apathy vs. Flow, Novak, Hoffman and Yung, 1997
 42. The Web stretches my capabilities to their limits  / I donít care much about it while I am searching the Web Apathy vs. Flow, Novak, Hoffman and Yung, 1997
43. I am active/ passive while I am searching the Web Apathy vs. Flow, Novak, Hoffman and Yung, 1997
44. I donít think/ I think about other things when I use the Web  Focused Attention, Ghani and Deshpande, 1994
45. I am soothed / I am alert while I am searching the Web  Relaxation vs. Arousal, Novak, Hoffman and Yung, 1997
46. I have patience/ no patience for mail or advertising while I am searching the Web
47. Web sites are reached easily/ Web sites are not reached easily
48. Web site address is clear / Web site address is not clear
49. The information that one can access with the Web is unlimited/ limited  Worry vs. Control, Novak, Hoffman and Yung, 1997
50. I trust / distrust all information in the Web sites
51. Commercial Web sites have easy / difficult billing option 8

7
7
7
7

6

6

6
5
5
5
5
4
4
4
4

4
4

4

3
3
3
3

3
3
 

3

3

3
3
3
3
3
3
 

3

2

2
2
2

2

2
2

2
 

2

2

2

2

2
2
2

2
2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Table 6. Surferís attribution factors of network navigation in CMEs

Factors and Loading Items Factor Loading     Item
  Mean Standard
Deviation
 Factor 1  (23.7% of variance)
"I am very skilled /not very skilled at using the Web"
"I consider/ donít consider myself knowledgeable about good Web search techniques"
"I am experienced / not experienced at Web searching"
"I know/ donít know my way around on the Net"
"I use precise search keywords/ I have poor choice of search words"
"I have good/ donít  have enough instructions on how to search"
"I know/ donít know the logical steps in Web searching"
"I have familiarity / no familiarity with the Internet"
"I understand/ donít understand computer language"
"I clearly know the right things to do/ feel confused about what to do while I am searching the Web"
"I know / donít know the exact Web address that I have to find"
"I get /donít get frustrated while I am searching the Web"
0.96
0.95

0.95
0.92
0.88

0.87

0.87
0.86
0.78
0.76

0.71

0.69
 2.96
 3.21

 2.68
 2.96
 2.83

 2.86

 2.94
 2.14
 3.59
 3.08

 3.82

 3.69
  2.25
  2.43

  1.96
  2.31
  2.13

  2.10

  2.15
  1.57
  2.71
  2.78

  2.85

  2.82
 Factor 2  (13.4% of variance)
"The linking of the computer is excellent/ terrible"
"Web sites are / are not reached easily"
"Readability of Web browsers is powerful/ not powerful"
"Web design are very real/ unreal"
"Most computers today have a good/ poor search engine"
"Navigating the Web with todayís Web browsers is natural/ unnatural"
"Web site address is clear/ not clear"
"Commercial Web sites today have broad/ narrow product ranges"
"The information that one can access with the Web is unlimited/ limited"
"The number of different ways one can interact with the Web today is not limited/ limited"
"It is easy/ not easy to find a bookmark that I am looking for"
0.79
0.78
0.77
0.75
0.73
0.71

0.70
0.68

0.66

0.64

0.59
 4.47
 4.38
 4.56
 4.42
 4.14
 4.18

 4.32
 5.28

 4.28

 5.52

 4.37
  1.76
  1.88
  1.86
  1.80
  2.13
  1.96

 2.04
 2.58

 2.24

 2.66

 2.23
 Factor 3    (7.9 % of variance)
"I am having fun searching the Web sites/ I get tired of Web searching"
"Using the Web provides/does not provide a good test of my skills"
"I think Web searching is an interesting/ boring job"
"Using the Web challenges/ does not challenge me"
"I am/ am not interested in finding products on Web sites"
0.78

0.77

0.73
0.56
0.51
 4.11

 4.39

 4.45
 5.07
 4.98
 2.71

 2.55

 2.46
 2.91
 2.94
 Factor 4    (6.4 % of variance)
"Modem is slow/ fast"
"The server of the computer is working well/ often down"
"Computer downloading takes reasonable time/ too much time"
"The computer is fast/ too slow"
"Web searching is too time-consuming/ not time- consuming"
0.83
0.77
0.75

0.73
0.48
 5.56
 5.18
 6.07

 5.55
 5.26
 2.24
 1.90
 2.33

 2.42
 1.94
 Factor 5    (5.7 % of variance)
"The Web stretches my capabilities to their limits / I donít care much about it while I am searching the Web"
"I enjoy the fast pace of the computer/ I feel apathetic while I am searching the Web"
0.97

0.96
 
 6.94

 5.66
 2.71

 2.45
 Factor 6     (5.1 % of variance)
"When I am using the Web, I am totally absorbed/ am not absorbed in what I am doing"
 "I donít think/ do think about other things when I use the Web"
0.77

0.76
6.21

6.68
 1.97

 1.88

Table 7. Attribution Variables in a Process of Network Navigation and Performance-Attribution-Expectancy Linkages
 

Performance on
Task A
Attribution  Impact on Expectancy Beliefs
 
Internal Stable - Skills, Challenges
  Positive
 
 
 
Internal Unstable - Focused Attention,
  Positive
                                 Flow
High
      Performance
 
External Stable - Interactivity (mapping,
  Positive
                              speed) at a general CME
 
 
External Unstable-Interactivity (mapping,
     No
                          speed) at a given time
                       during the Activity/Survey    impact
 
 
Internal Stable - Skills, Challenges         Negative
 
 
 
Internal Unstable - Focused Attention,
  Positive
                                  Flow
Low
         Performance
 
External Stable - Interactivity (mapping,        Negative
                              speed) at a general CME
 
 
External Unstable-Interactivity (mapping,
    No
                                speed) at a given time
                       during the Activity/Survey    impact