Designing Business-To-Consumer (B2C) Interface Metaphors:



Designing Business-To-Consumer (B2C) Interface Metaphors:

An Empirical Investigation

Abstract

The emergence of electronic commerce has pushed information technology to an increasingly heterogeneous set of users (e.g., customers) who interact with a wide variety of user interfaces. As a result, the need for user-friendly, intuitive interfaces has become an urgent issue in electronic commerce. Interface metaphors are a popular means for facilitating user interaction via familiar, concrete objects or attributes. This research explores the use of concrete attributes derived from the physical business domain as a technique for designing business-to-consumer (B2C) interface metaphors. A laboratory experiment was designed to test the effectiveness of a concrete interface metaphor for presenting both textual and graphical information, as compared to an abstract interface metaphor. The retention/recall of the information was measured for two different interface metaphors, with subjects being tested both the day of the treatment and after a two-day lag. Results revealed that the concrete interface metaphor stimulated higher levels of retention and recall of information, particularly for customers who possess a weak mental model of the business domain. In addition, it was observed that the concrete interface metaphor stimulated a higher level of graphical information retention without adversely affecting the retention of textual information.

Subject Areas: Cognitive Fit, Customer Interaction, Electronic Commerce, Information Presentation, Interface Design, Mental Models, Metaphor

1. Introduction

The emergence of electronic commerce presents organizations with an unprecedented opportunity for interacting with its customers[1] via information technology (IT). Yet, organizations continue to struggle with harnessing the power of business-to-consumer (B2C) electronic commerce. A primary challenge is realizing that the consumer, for all practical purposes, is fundamentally different from the traditional end-user (Koufaris, 2002). The traditional system user typically utilizes IT within a job-related context to facilitate a relatively homogenous set of tasks. Conversely, on average, consumers are a less technical set of users (Pitkow & Kehow 1996) who find themselves inundated with information. They utilize IT on a relatively infrequent basis to perform a variety of tasks such as information gathering and product/service purchasing. The emergence of the consumer as an end-user stimulates the need to design and implement highly intuitive interfaces that effectively present information to support an increasingly heterogeneous set of tasks. Therefore, novel approaches to interface design are necessary to fully realize the potential of B2C electronic commerce.

The ability to design intuitive user interfaces has been an ever-present challenge. A common method for creating such interfaces is through the use of interface metaphors. Metaphorical interfaces have been employed at various points along the evolutionary path of information technology (IT). The use of metaphorical interfaces began with the spreadsheet Visicalc in the late 1970's and fully manifested itself when Apple introduced the desktop metaphor as part of its graphical user interface (GUI) for Lisa and the Macintosh (Raskin, 1997). These interface metaphors proved to be effective because they presented users with tangible, concrete, recognizable objects that were used to facilitate their understanding of the computer application.

In B2C electronic commerce, a primary goal is the ability to present a virtual representation of products and/or services. Because metaphors are defined by a user’s perception of objects in his/her environment (Lackoff & Johnson, 1980), business domains are a viable source for deriving concrete interface metaphors. For instance, these domains can include the product domain (e.g., graphical representation of a product) or the service domain where products/services are consumed (e.g., restaurant). By creating an interface metaphor based on the attributes from the business domain, organizations can present consumers a more navigable and intuitive interface.

While academicians in the area of human-computer interaction (HCI) have been quite active in exploring the issues surrounding interface metaphors, most of this research has addressed conceptual (Carroll & Thomas, 1982; Carroll & Mack, 1985; Dieberger & Frank 1998), methodological (Madsen, 1994), and usability issues (Erickson, 1990; Kay, 1990). While some empirical research exists (Golovchinsky & Chignell, 1997; Vaughan, 1998; Kim, 1999), it hasn’t focused on the use of interface metaphors in business applications. The emergence of electronic commerce presents an unprecedented opportunity to better understand how organizations can utilize interface metaphors to more effectively interact with consumers. Thus, the purpose of the research presented in this paper is to gain a better understanding of how consumers perceive and interpret different interface metaphors within an electronic commerce context.

The literature for this study comes from a number of different sources including metaphor theory, interface metaphor design/usability, mental model theory, and cognitive fit theory. Using this literature as a theoretical foundation, specific hypotheses are proposed that explore how consumers interpret interface metaphors. A laboratory experiment was conducted in an effort to test these hypotheses. An analysis of the results is presented. Finally, contributions, limitations, and future work for this type of research are discussed.

2. Interface Design for Electronic Commerce: Theoretical Foundations

A number of interface design challenges have emerged with the growth of the WWW and related electronic commerce applications, particularly online marketing activities (Lohse & Spiller, 1998). An examination of customer attitudes and expectations (Jarvenpaa & Todd, 1996) has led to the identification of web usability criteria (Chau et al., 2000; Nielsen, 2000) and, subsequently, the validation of instruments to measure these criteria (Agarwal & Venkatesh, 2002; Palmer, 2002). From these studies, a key issue that consistently emerges is the need for meaningful and accessible content. Or simply put – effective information presentation.

The effect of information presentation on how consumers’ store and process information has been the topic of numerous marketing studies, particularly as it relates to consumer decision-making (Bettman & Kakkar, 1977; Bettman, 1979; Johnson & Russo, 1984). It has been argued that electronic commerce represents the intersection between the information systems and marketing disciplines (Zinkhan & Watson. 1998), providing opportunities to apply marketing principles to the IS domain. Contrary to the traditional marketplace, electronic commerce is conducted over a digital medium that is inherently abstract and, as a result, creates a significant communication gap between system designers and consumers. Bridging this gap, typically through the utilization of user interfaces, has been a key issue for supporting consumer information processing with electronic commerce applications.

The user interface has been defined as "software that shapes the interaction between user and computer" and that "serves as a kind of translator, mediating between two parties, making one sensible to the other" (p. 14) (Johnson, 1997). The ability to design an effective user interface becomes an even more complex challenge when a customer assumes the role of the end-user. User characteristics for customers are different from the traditional dedicated user (Pitkow & Kehoe 1996) as customers are generally less computer literate and their interaction domain varies to a much greater degree. Taking these issues into account, the need for intuitive customer interfaces becomes very apparent. A proven technique for designing an intuitive interface is the user-centered perspective (Norman & Draper, 1986), or for electronic commerce, the customer-centered approach (Beyer & Holtzblatt, 1998).

An effective method for creating a user-centered interface is through the use of interface metaphors (Erickson, 1990). Interface metaphors are particularly conducive to effective user-centered interface design because of their ability to leverage recognizable, concrete objects (e.g., folders, trash cans) to facilitate a user’s understanding of a software system (e.g., Macintosh operating system). As a result, the user perceives the interface to be more intuitive and, subsequently, more effective in terms of certain usability criteria (e.g., accuracy, information retention, etc.).

Currently, the de facto interface design technique for the WWW and, in particular, electronic commerce uses “frames” to organize logical information categories. The frame interface metaphor is useful for certain contexts, particularly the presentation of information in logical categories. Yet, information categories are inherently abstract within this type of metaphor as they are typically represented as textual or graphical hyperlinks. From an interface metaphor perspective, these links have no concrete meaning in the customer’s perception of the physical business domain. Thus, the excessively abstract nature of the frame interface metaphor along with increasing heterogeneity of both users and business domains has stimulated a need to identify alternative interface metaphors that are more customer-centered.

To understand how interface metaphors can be designed to more effectively present information to customers, several conceptual areas must be explored and understood. First, the use of interface metaphors as the foundation for an effective electronic commerce interface is discussed. Second, the customer’s domain familiarity (i.e., his/her perception of the interaction domain) is presented as a key factor for effective interface design. Finally, because interface metaphors present both textual and graphical information, Cognitive Fit theory is used as a foundation for understanding how a user’s domain familiarity affects how he/she perceives these two types of information.

2.1 Interface Metaphors

Lackoff and Johnson (1980) state, "the essence of metaphor is understanding and experiencing one kind of thing in terms of another" (p. 5). Metaphors are human derived models that apply tangible, concrete, recognizable objects (i.e., source domain) to abstract concepts and/or processes (i.e., target domain) (Figure 1) (Baecker, et al., 1995). A good illustration is the analogy "argument is like war" where war is the source domain and argument is the target domain (Lackoff & Johnson, 1980). The fundamental premise of metaphors is to facilitate the comprehension of an intangible, abstract domain (e.g., argument) via the attributes of a more tangible, concrete domain (e.g., war). The role of metaphor in the design of information systems has been approached from a number of different perspectives. It has been viewed philosophically (Coyne, 1995; Turkle, 1995) as well as pragmatically (Johnson, 1994; Madsen, 1994; Raskin, 1997). It has been examined at the individual level (Erickson, 1990; Kay, 1990) and also at the organizational level (Kendell & Kendell, 1993). When viewed from an individual and pragmatic perspective, metaphors are typically perceived as a user interface design technique (Erickson, 1990; Madsen, 1994). The research presented in this paper primarily focuses on the use of metaphors for interface design and how it affects individuals (i.e., customers).

Early uses of interface metaphors adhered to the basic tenets of metaphor theory by using concrete, recognizable objects from the user domain with the flagship example being the desktop metaphor. The use of interface metaphors in electronic commerce is a bit more complex because a higher degree of variability exists not only with respect to the system domains (i.e., software interfaces), but also with the types of users (i.e., customers). As stated earlier, the WWW has been used extensively for information presentation. This led to the widespread acceptance of the frame interface metaphor, which organizes information into logical categories. Information categories are represented by hypermedia links that have no tangible or concrete meaning to the customer.

***** Insert Figure 1 Here *****

By using abstract objects as the basis for an interface metaphor, one runs the risk of violating many of the basic interface metaphor design principles (Madsen, 1994). These principles recommend using tangible, concrete, recognizable objects (Baecker, et al., 1995) that reside within the interaction domain (Carroll, et al. 1988). According to Erickson (1990), the use of such objects creates an interface metaphor that contains a structure that is visually representative of the interaction domain that facilitates a user’s understanding of the interface. Because the information being presented to customers in an electronic commerce interface is often related to their perception of the physical business domain (e.g., products, service environments, etc.), these domains act as potential sources for deriving interface metaphors. Hence, the customer’s perception of the interaction domain (i.e., physical attributes from the business domain) can be used to derive interface metaphors that effectively present information to customers (see Figure 1). In turn, more effective information presentation allows customers to retain and recall information for a number of consumer-related tasks, particularly decision-making (Johnson & Russo, 1984). Thus, we propose the following hypothesis:

Hypothesis 1: An interface metaphor that is based on concrete business domain attributes will enable a user to retain significantly more information when compared to an interface metaphor that is based on abstract business domain attributes.

An interface’s degree of usability is often user-dependent. In other words, individual differences exist that dictate the effectiveness of an interface metaphor. One such difference is that a customer’s prior experience with a particular business domain may affect how he or she interprets an interface metaphor that is based on the same domain. This dynamic can be better understood by examining a customer’s familiarity with the business domain, which is discussed in the following section.

2.2 Domain Familiarity

Domain familiarity can be explained using mental model and schema theories (Rumelhart & Norman, 1978; Rumelhart, 1980). Schemata consist of knowledge structures that act as means to store concepts in human memory. Schema are linked to related sub-schemata that, as a whole, represent an individual's knowledge about a particular domain (Rasmussen, 1990) and is considered a key component to the creation of mental models (Norman, & Draper, 1986). Norman (1990) frames the concept of mental models into three different areas: the design model, the user's model, and the system image (See Figure 2). Further, the design model is a conceptual model that the designer possesses of the system. The system image is the actual physical structure of the system. Finally, the user model is the user’s conceptual view (or mental model) of the system. Ideally, mental model theory stresses the importance for the design and user models to be tightly coupled (Norman & Draper, 1986). Simply stated, the system designer must be able to create functionality that is congruent with the user’s view of the system, making the user’s perception of the business domain an important design consideration and, for the purpose of this research, an instantiation of a user’s mental model.

***** Insert Figure 2 Here *****

The relative strength of a user’s domain familiarity can be further understood when viewed from an expert/novice perspective. Expert/novice studies have been approached from a number of non-IT related domains such as chess (Chase & Simon, 1973) and the analysis of physics problems (Chi, et al., 1981). In these studies, experts were able to create high levels of abstractions to support their problem-solving skills while novice users tended to focus on literal and concrete aspects of the problem domain. These same cognitive processes can be used to further understand how different users interpret interface metaphors.

Considerable attention has been focused on how experts and novices interpret metaphors. Gillan, et al., (1995) examine how experts and novices react to interface metaphors from both computing experience and cognitive ability perspectives. Results of this study demonstrated that experts possessing computer experience could interpret metaphors using abstract attributes. At the same time, novice users were observed to interpret metaphors by using physical or concrete attributes. Gentner (1988) demonstrated that experienced users not only handle highly abstract metaphors, but they are also able to interpret concrete metaphors effectively. Gentner goes on to explain that novice users, because of their limited ability to interpret abstractions, need metaphors that are based on functionality and are concrete in nature. Therefore, we offer the following hypothesis for understanding the relative strength of a user’s domain familiarity as it relates to retaining and recalling information presented in a user interface:

Hypothesis 2: A user who possesses strong familiarity of the business domain will retain significantly more information than a user who possesses weak familiarity of the business domain.

This research makes a direct comparison between the key components of metaphor theory (i.e., source and target domains) and mental model theory (i.e., design model, system image, and user model). It is our contention that a parallel should be drawn between the user model in mental model theory and the proposed inclusion of the customer’s perception of the business environment as a source domain in metaphor theory (see Figures 1 and 2). While mental model theory helps explain some of the possible implications of how the user’s familiarity of the application domain can affect how they interact with an interface metaphor, the type of information presented within the interface needs to be addressed as well. The following section uses cognitive fit theory as a foundation for anticipating some of the issues related to how information is presented in an interface metaphor.

2.3 Information Presentation

How textual and graphical information are represented in an interface metaphor presents a design challenge. Cognitive fit theory suggests that there is a strong link between how information is represented in an interface and the effectiveness with which users interpret/use the information (Vessey, 1991; Dennis & Carte, 1998; Swink & Speier, 1999). Vessey and Galletta (1991) compared how users utilized symbolic (i.e., textual) and spatial (i.e., graphical) information to better support problem-solving tasks. Several researchers have contributed to the distinction between these information acquisition tasks (Washburne, 1927; Umanath & Scammell, 1988; Umanath, et al., 1990). Symbolic tasks are more specific (i.e., discrete) instances of data and don’t need to be linked to other pieces of data in order to be meaningful. Spatial tasks are viewed from a holistic perspective and imply the presence of relationships between specific pieces of data. A key insight from Vessey and Galletta’s (1991) study was that problem-solving was enhanced when there was a close coupling between the task type and type of information used to support the task. With respect to interface design, this raises the issue of how graphical and textual information should be presented to the user. Since most user interfaces, particularly in the electronic commerce arena, are graphical in nature, the presentation of spatial information is very common. However, the need to incorporate symbolic information within a graphical user interface (GUI) is still an obvious requirement.

Cognitive Fit Theory has focused on how symbolic and spatial information presentation affects specific types of problem-solving or decision-making tasks. Task types in electronic commerce can range from very specific (e.g., pricing for specific product/service) to very broad (e.g., researching product/service offerings). For a broader range of tasks, particularly related to electronic commerce, the symbolic information is invariably coupled with spatial information. Ideally, an interface metaphor should be designed that facilitates a customer’s ability to retain and recall BOTH symbolic and spatial information. Yet, the significance of Cognitive Fit Theory for designing interfaces remains very important, particularly the need to qualify the differences between symbolic and spatial data when evaluating new techniques for information presentation. Thus, the following research hypotheses are proposed a priori to examine the relationship between a customer’s domain familiarity (i.e., strong vs. weak) and the type of interface metaphor (i.e., concrete vs. abstract) along with any resulting distinction between the retention of symbolic versus spatial information.

2.3.1 Symbolic Information: Symbolic information is the overt presentation of text and, by its nature, is often organized by logical categories. As stated earlier, the frame interface metaphor is a common means for presenting these categories, which typically use abstract objects (e.g., hyperlinks) to present symbolic information. For a focused task (e.g., information search), Cognitive Fit Theory supports the use of a frame interface metaphor for presenting symbolic information. Taking into account the broader nature of electronic commerce tasks, we look to the interface metaphor and mental model literature to possibly extend Cognitive Fit theory and provide an understanding of how symbolic information can be effectively presented in concert with spatial information.

As explained by Gentner (1988), users who possess a strong familiarity of the problem domain have an ability to create abstractions that allow them to interpret and understand metaphorical representations. Therefore, users with a strong familiarity of the business domain can successfully interpret information that is presented via an interface metaphor based on either concrete or abstract attributes. Gillan, et al. (1995) clarified some key differences between users who possess strong domain familiarity (SDF) and those who possess weak domain familiarity (WDF). Most notably, SDF users can effectively interpret both abstract and concrete metaphors while WDF users prefer a concrete metaphor. That is, if an interface metaphor is based on abstract attributes (e.g., frame interface metaphor), users who possess a WDF level of domain knowledge will have more difficulty interacting with the interface as compared to a SDF. Thus, we offer the following hypothesis that a user with strong familiarity of the business domain can effectively retain and recall information presented via an abstract interface metaphor.

Hypothesis 3a: When using an interface metaphor based on abstract business domain attributes, the amount of symbolic information retained by SDFs will be significantly MORE than the amount of symbolic information retained by WDFs.

While we contend that users with a weak familiarity of the business domain are less effective when interpreting an abstract interface metaphor, WDF users are more effective when interacting with concrete interface metaphor (Norman & Draper, 1986; Gillan, et al., 1995). Concrete objects, as compared to the abstract domain, will build a stronger mental model for the WDF user and, therefore, enhance their ability to retain more symbolic information.

Hypothesis 3b: The amount of symbolic information retained by WDFs when using an interface metaphor based on concrete business domain attributes will be significantly MORE than the amount of symbolic information retained by WDFs when using an interface metaphor based on abstract business domain attributes.

2.3.2 Spatial Information: Spatial information is the presentation of information in the form of graphical objects. A concrete interface metaphor uses graphically oriented objects from the interaction domain while a frame interface metaphor is typically based on abstract information categories. Because of its graphical nature, an interface metaphor based on graphical concrete attributes is a better match for presenting spatial information (Vessey, 1991; Vessey & Galletta, 1991). As was the case with the preceding hypotheses (i.e., 3a and 3b), our goal is to understand how to effectively present spatial information.

According to mental model theory (Norman & Draper, 1986), the interaction with the familiar concrete objects will build a stronger mental model that is more aligned with the actual business domain, resulting in a higher degree of domain familiarity. When presented with an interface that is devoid of such concrete objects, the SDF user can use their domain familiarity to overcome the ambiguity of an abstract interface metaphor (Gentner, 1988) while the exposure to a relatively more abstract interface metaphor along with a weaker domain familiarity will make the WDF user less effective (Gillan et al., 1995).

Hypothesis 4a: When using an interface metaphor based on abstract business domain attributes, the amount of spatial information retained by SDFs will be significantly MORE than the amount of spatial information retained by WDFs.

The effect of a concrete interface metaphor on a WDF user, as compared to a SDF user, can be stronger because the WDF user does not have an existing mental model that can potentially conflict with the interpretation of the interface metaphor (Adleson, 1984). The interaction with the familiar concrete objects that exist in the business domain interface will allow the WDF user to create a stronger mental model than by interacting with objects that are perceived to be more abstract (Norman & Draper, 1986). This allows the WDF user to leverage their improved mental model to retain more information about the business domain (Rumelhart, 1980).

Hypothesis 4b: The amount of spatial information retained by WDFs when using an interface metaphor based on concrete business domain attributes will be significantly MORE than the amount of spatial information retained by WDFs when using an interface metaphor based on abstract business domain attributes.

For a SDF user with an existing mental model that corresponds to a strong familiarity of the general business domain, the interaction with a concrete interface metaphor allows for higher levels of information retention (Rumelhart & Norman, 1978).

Hypothesis 4c: The amount of spatial information retained by SDFs when using an interface metaphor based on concrete business domain attributes will be significantly MORE than the amount of spatial information retained by SDFs when using an interface metaphor based on abstract business domain attributes.

4. Method

***** Insert Figure 3 Here *****

***** Insert Figure 4 Here *****

The hypotheses presented in the previous section were proposed to test both the main effects (see Figure 3) related to information retention as well as simple main effects (see Figure 4) that qualify the type of information retained (i.e., symbolic vs. spatial). An experiment was conducted to gain insight into the most effective method for designing an interface that maximizes the user’s ability to retain both spatial and symbolic information. The independent variables were mode of interface and domain familiarity. The dependent variable consisted of a measure for user information retention.

4.1 Independent Variables

4.1.1 Mode of Interface: Users communicate with computers via two languages: action and presentation (Gerlach & Kuo, 1991). Users provide the action language that comes in the form of requests for actions to be performed by the application. The computer, upon receiving a request by a user, utilizes presentation language to provide feedback or to present information. Two front-end interfaces were designed to present the exact same information, in both a symbolic and spatial format (see Exhibits 1(a) and 1(b) in the Appendix). To ensure that the same information was presented in both interfaces, an information repository was created to store the text and graphics that were used in each interface. An information inventory checklist was used to make sure that both text and graphics from the central repository were included in each interface. These interfaces encompass the mode of interface independent variable. The primary difference in the systems was how the two interfaces structured the presentation language.

The design principles presented by Erickson (1990) and Madsen (1994) were referenced to create a metaphorical interface that incorporates objects and processes from the designated business domain. Therefore, tangible, concrete, recognizable objects from the business domain were used in the source domain to facilitate the user’s understanding of the target domain (Baecker, et al., 1995). To illustrate, an interactive map of the business was designed that would allow the user to click on building structures that would, in turn, present context-specific information.

The other interface was based on an abstract metaphor and was designed using established HTML design principles that are commonly used in industry (Vora, 1998). This particular interface used a frame metaphor that is commonly used for web interfaces. The same graphics and text were used in both interfaces in this study. The primary difference in the frame metaphor, when compared to the concrete interface metaphor, was that information was organized by abstract categories (i.e., logical information group) instead of the physical business domain.

4.1.2 Domain Familiarity: This independent variable was included because prior research points to the effect that a person’s interpretation of the problem domain has the ability to interpret a metaphor effectively (Gentner, 1988). Using the concept of mental models as a theoretical basis, subjects were given a questionnaire to determine their degree of domain familiarity and were assigned to one of two groups: strong or weak. Norman (1990) defines a mental model as “models people have of themselves, others, the environment, and the things with which they interact” and states that “people form such models “through experience, training, and instruction” (p. 17). Consistent with prior mental model research (Satzinger & Olfman, 1998; Shayo & Olfman, 2000; Smith-Jentsch, et al. 2001), this variable was operationalized by testing the accuracy of a subject’s mental representation of the problem domain. A questionnaire was designed to assess a subject’s past interactions with the business domain and measure the accuracy of his/her mental representation. Each question was weighted equally. The questionnaire contained both correct and incorrect descriptive features that related to the problem domain. If a potential subject correctly selected a feature, they received a certain number of points. If they incorrectly selected a feature, they had a certain number of points subtracted from their total. After subjects completed the questionnaire, a quantitative score was calculated. Scores were placed into three groups: high, medium, and low. The medium group was eliminated from this study in an effort to polarize the two domain familiarity groups.

4.2 Dependent Variable

The dependent variable for this study was a measurement of the amount of information the user retained (and recalled) as a result of the mode of interface. Information retention was chosen because it is considered to be a primitive, fundamental requirement for consumer information processing (Johnson & Russo, 1984), particularly as it relates to information presentation and decision-making (Bettman & Kakkar, 1977; Bettman, 1979). From an IT research perspective, the use of information recall has been used in past information presentation studies, particularly the tables vs. graphs literature (Umanath & Scamell, 1988; Umanath, et al., 1990).

4.3 Problem Domain

The problem domain for this study was a hypothetical vacation resort that was based on typical resort businesses. The use of a hypothetical vacation resort was attractive because the possibility of a subject having specific domain knowledge about a particular resort was eliminated. However, the ability to test the domain familiarity effect still exists because subjects can have either a lot or very little experience with vacation resorts in general. Currently, most resort organizations present customers with various forms of information on the Internet including room rates, recreational activities, reservations, etc. The typical interface metaphor for this type of information presentation is a hypermedia approach using frames. Given the goal of this research, two types of interfaces were developed, the typical abstract interface metaphor approach and a concrete interface metaphor that utilizes tangible objects (e.g., buildings and recreational areas) from the business domain. Users were presented with a very broad information-browsing task (i.e., learn as much about this resort as possible) using one of the two interface types. This information-browsing task was intentionally designed to be broad in an effort to expose the user on both symbolic and spatial information.

4.4. Instrument Validation

Two pilot studies were administered to test the experimental procedures and validate the research instrument. Thirty-nine subjects participated in the first pilot study. Results from the first pilot study were used to refine the instrument and to assess both the content validity and reliability of the measurement instrument.

The measurement instrument was carefully designed in an effort to maximize content validity. An information inventory was created that organized all the information related to the problem domain into two categories: symbolic and spatial. The information inventory was referenced to obtain a broad range of questions and, hence, ensure that instrument contained adequate breadth of scope. In many cases, content validity can be verified by allowing an expert to examine the instrument (Huck & Cormier, 1996). The instrument was given to colleagues who possessed a strong working knowledge of instrument creation as well as the problem domain. After they carefully reviewed it, their suggestions were used to strengthen its validity. Once the measurement instrument was finalized, it was administered to a set of subjects for the purpose of further ascertaining its content validity. While the strong and weak domain familiarity subjects were exposed to the two interface types in the pilot study, the moderate domain familiarity group (i.e., the group that was thrown out in order to create the polarity between the strong and weak mental model types) was used to test content validity. Without being exposed to either interface, fourteen subjects were asked to complete the questionnaire. The average score was below one point (.58 out of 30) and most of the answers that were correct were the product of random guessing, thus demonstrating that the subjects were limited in their ability to intuitively answer the questions. However, there was some concern about two of the questions because a subject could intuitively ascertain the correct answer. Therefore, these questions (one symbolic and one spatial) were eliminated to minimize this effect. The final version of the instrument contained 30 questions: 15 symbolic and 15 spatial. Overall, these tests provided sufficient evidence that content validity for the measurement instrument was sound.

A second pilot study was administered for the purpose of testing the reliability of the measurement instrument. Twenty-two subjects participated in the second pilot study. Once again, Pearson's product-moment correlation was used to test for the reliability of the measurement instrument. Because the majority of hypotheses in this study focused on information sub-categories (i.e., symbolic vs. spatial information), it was necessary to ensure the reliability of a subject's retention of these sub-categories across the two questionnaire forms. After subjects interacted with the concrete interface metaphor, they were randomly assigned to either questionnaire Form A or B. The average score for each sub-category was compared across the two form types, with a correlation coefficient of .92.

4.5 Experiment

The subjects for this study were undergraduate students who were enrolled in an introductory management information systems course at a major university. A total of eighty-eight subjects (twenty-two in each treatment group) participated in the experiment. A questionnaire was administered to measure the subject's information retention. After subjects were categorized by domain familiarity, either weak or strong, they were randomly assigned within each group to either a concrete or abstract interface metaphor. The information presented in these interfaces was categorized into two areas: symbolic and spatial. Textual and graphical information were measured separately to make a distinction between symbolic and spatial information types. Lastly, focus groups were conducted in an effort to gather qualitative data to support this study and identify future research opportunities.

Because information retrieval is affected over time, typically attributed to interference (Jenkins and Dallenbach, 1924), it was deemed necessary to test for information retention after a certain period of time. Ashcraft (1995), using the Ebbinghaus (1885/1913) forgetting curve, illustrates that retention is significantly affected over a two-day period. Thus, after a two-day lag, a different instrument (i.e., the other form) was administered to test the persistence of the information retention. The instrument contained information that was similar in format and complexity, but different in context in an effort to avoid a learning effect. This instrument was tested for reliability against the first instrument during pilot tests, which was discussed in the Instrument Validation section.

5. Analysis of Results

The overall average retention scores were 15.11 and 9.61 (re-test)[2], out of a total of 30 questions. The subsequent discussion of the results from this study is organized by research hypotheses. Tables 1 and 2 present the ANOVA results for the overall information retention scores for both the same-day and re-test instruments. Table 3 summarizes the overall averages for information retention in this study.

***** Insert Table 1 Here *****

***** Insert Table 2 Here *****

***** Insert Table 3 Here *****

The research hypotheses are organized across four logical categories: 1) mode of interface, 2) domain familiarity, 3) symbolic information retention, and 4) spatial information retention. Hypotheses 1 and 2 tested the main effects for each of the independent variables: mode of interface and domain familiarity. Hypotheses 3 and 4 tested both symbolic and spatial information retention across treatment groups. This study consisted of four treatment groups: SDFs who interacted with the concrete interface metaphor (denoted as SC), SDFs who interacted with the abstract interface metaphor (denoted as SA), WDFs who interacted with the concrete interface metaphor (denoted as WC), and WDFs who interacted with the abstract interface metaphor (denoted as WA). Hypotheses 3 and 4 were broken down into sub-hypotheses (e.g., H3a), which allowed for the comparison of individual treatment groups (e.g., SC compared to WC). The pairwise comparisons were calculated using Tukey's method, which is considered more conservative than other popular techniques such as Fisher's LSD (Huck & Cormier, 1996).

5.1 Mode of Interface (H1)

This hypothesis was designed to test the effect that mode of interface had on a subject’s ability to retain and retrieve information. For the 1st test and re-test, respectively, the average score for subjects in the concrete interface metaphor group was 17.30 and 11.87 while the average score for subjects in the abstract interface metaphor group was 12.93 and 7.36. For both tests, a significant difference was found between the volume of information retained by subjects in the two mode of interface groups (p = .0001).

5.2 Domain Familiarity (H2)

This hypothesis was designed to test the effect that a subject’s domain familiarity had on his/her ability to retain and retrieve information. For the 1st test, the average score for subjects in the strong domain familiarity group was 15.59 and the weak domain familiarity group was 14.64, which did not reveal a significant difference (p = .2386). Conversely, the re-test produced a wider gap with the strong domain familiarity group averaging 10.55 as compared to the weak domain familiarity group at 8.68. Thus, relative significance (p = .0215) was observed in the re-test.

Also, there was no consistent evidence of interaction effects between the two independent variables for either the 1st test (p = .0963) or the re-test (p = .2334). The following hypotheses (H3 and H4) provide more detailed insight into the simple main effects between these two variables by testing the mean differences across specific treatment groups (see Table 4)[3].

***** Insert Table 4 Here *****

5.3 Symbolic Information Retention (H3a, H3b)

These hypotheses were designed to test how effectively subjects retain symbolic information. An analysis of the data indicated that a significant difference did not exist between any of the four groups. Table 5 shows a more detailed breakdown of the comparison of the individual mean values.

***** Insert Table 5 Here *****

5.2.1. Hypothesis 3a: This Hypothesis postulated that SDFs who interacted with the abstract interface metaphor would retain more symbolic information than WDFs who interacted with the abstract interface metaphor. For the 1st test and re-test, respectively, the averages for the SA group were 7.36 and 4.64 with the averages for the WA group being 6.86 and 3.86. Based on the analysis of the data, the results failed to reject the null hypothesis. More specifically, the intersection of the SA and WA groups in Table 5 indicates no significance for either the 1st test (p = .846) or the re-test (p = .506).

5.2.1. Hypothesis 3b: This Hypothesis postulated that WDFs who interacted with the concrete interface metaphor would retain more symbolic information than WDFs who interacted with the abstract interface metaphor. For the 1st test and re-test, respectively, the averages for the WC group were 7.54 and 5.00 with the averages for the WA group being 6.86 and 3.86. Based on the analysis of the data, the results also failed to reject the null hypothesis. Further, the intersection of the WC and WA groups in Table 5 indicates no significance for either the 1st test (p = .682) or the re-test (p = .178).

5.3 Spatial Information Retention (H4a, H4b, H4c)

These hypotheses were designed to test how effectively subjects retain spatial information. Data analysis indicated that a significant difference existed between at least two of the four treatment groups. Table 6 shows a more detailed breakdown of the comparison of the individual mean values.

***** Insert Table 6 Here *****

5.3.1. Hypothesis 4a: This Hypothesis postulated that SDFs who interacted with the abstract interface metaphor would retain more spatial information than WDFs who interacted with the abstract interface metaphor. For the 1st test and re-test, respectively, the averages for the SA group were 6.73 and 4.14 with the averages for the WA group being 4.91 and 2.09. Based on the analysis of the data, the results (see Table 6) do not support the rejection of the null hypothesis for the 1st test (p=.187). While the results from the re-test are more encouraging, only moderate support for rejecting the null hypotheses was found (p=.062).

5.3.2. Hypothesis 4b: This Hypothesis postulated that WDFs who interacted with the concrete interface metaphor would retain more spatial information than WDFs who interacted with the abstract interface metaphor. For the 1st test and re-test, respectively, the averages for the WC group were 9.95 and 6.41 with the averages for the WA group being 4.91 and 2.09. Based on the analysis of the data, the results warrant the rejection of the null hypothesis. More specifically, the intersection of the WC and WA groups in Table 6 indicates significance for both the 1st test (p = .000) and the re-test (p = .000).

5.3.3. Hypothesis 4c: This Hypothesis postulated that SDFs who interacted with the concrete interface metaphor would retain more spatial information than SDFs who interacted with the abstract interface metaphor. For the 1st test and re-test, respectively, the averages for the SC group were 10.23 and 7.09 with the averages for the SA group being 6.73 and 4.14. Based on the analysis of the data, the results warrant the rejection of the null hypothesis. More specifically, the intersection of the SC and SA groups in Table 6 indicates significance for both the 1st test (p = .001) and the re-test (p = .002).

6. Summary of Results

Three of the seven hypotheses (H1, H4c, H4b) in this study were strongly supported in the results of the 1st information retention questionnaire. For the re-test questionnaire, four of the seven hypotheses (H1, H2, H4b, H4c) were strongly supported with H4a receiving moderate support. A summary of these results is presented in Table 7.

***** Insert Table 7 Here *****

The lack of significance with respect to the retention of symbolic information was attributed to two factors. First, it appears that the benefits of interacting with the concrete interface metaphor may have been neutralized by the strong cognitive fit of the information categories presented in the frame interface metaphor. Because the abstract (i.e., frame) interface metaphor organized the symbolic information into categories, it can be postulated that this organization technique provides a more effective cognitive fit for presenting and retaining textual information. Interestingly, the retention of symbolic information was not significantly affected when subjects interacted with the concrete interface metaphor (see Figure 5). Further research needs to be conducted that more clearly isolates the cognitive fit effect from the effect of the concrete interface metaphor.

***** Insert Figure 5 Here *****

***** Insert Figure 6 Here *****

The results associated with spatial information strongly supported the majority of these respective hypotheses. Two factors were of primary importance. First, the degree of interactivity in the concrete interface metaphor placed an impetus on user interaction with the concrete, graphical objects in the problem domain. The concrete interface metaphor used these objects as the primary means of interaction, thus creating a strong mental model of the spatial information in the problem domain. Subjects then utilized this mental model to retain and recall significantly more spatial information than subjects who interacted with the abstract interface metaphor (see Figure 6). This effect was explained using metaphor, mental model, and cognitive fit theories. These theories apply to both subjects with strong and weak domain familiarity. The use of concrete objects in an interface metaphor allowed a subject with weak domain familiarity to build a stronger mental model. Subjects who possessed strong domain familiarity were able to interact with the concrete objects as well. Therefore, both groups of subjects were able to retain significantly more spatial information when interacting with the concrete interface metaphor.

The most compelling observation is that while there was not a significant difference with respect to symbolic information retention, the concrete interface metaphor did not adversely affect a subject's ability to retain this type of information. Yet, at the same time, subjects were able to retain more spatial information by using the concrete interface metaphor. Therefore, the concrete interface metaphor appears to be a viable design technique for displaying both symbolic and spatial information.

6.1 Contributions

This experimental study made both theoretical and practical contributions to the field of Information Systems. The following section expands upon these contributions.

6.1.1 Theoretical Contributions: This study makes contributions by integrating theory from two research areas: human-computer interaction and cognitive psychology.

The area of human-computer interaction focuses on the design and use of interface metaphors. Interface metaphors use metaphor theory as a basis for understanding application domains. The source domain is used as a reference point for what the interface metaphor should encompass and how it should facilitate the understanding of the target domain. By using the business domain as a source for identifying interface metaphors (see Figure 1), the methodology for creating these types of interfaces has been considerably altered. The inclusion of the business domain affects the three key areas of interface metaphor development: generation, evaluation, and development (Madsen, 1994). Hence, looking to the customer’s perception of the business domain as the foundation for the source domain of the interface metaphor adds structure to existing design methodologies.

Also, the results indicate that the concrete interface metaphor provides a more effective presentation language for the interface while keeping the action language consistent. The action language is simply how a user requests action on the part of the computer interface. Both interfaces provided this capability via the mouse and the functionality of the web browser. The major difference was how the presentation language was structured: concrete versus abstract interface metaphor. Particularly with respect to spatial information, the results indicated that the concrete interface metaphor provided subjects with a more effective means to retain information about the problem domain.

The contributions to cognitive psychology are three-fold. First, mental model theory and its relationship to domain familiarity were used to justify the inclusion of the business domain for the design and development of interface metaphors. Norman (1990) stressed the importance of coupling the user and design models when designing a system. This principle directly supports the coupling of the business and source domains in interface metaphors. Because the results indicated a statistically significant difference between the concrete and abstract interface metaphors, Norman's mental model theory was reaffirmed in the process.

The second theoretical area that should be discussed is cognitive fit theory. Vessey’s (1991) theory suggested that information should be presented to users in a manner that is congruent with the task. Previous research pointed to several task types such as problem-solving, decision-making, and software application design. This study contributes evidence for the inclusion of another task type: information search/browsing. Additionally, these results point to the possibility that a concrete interface metaphor, when measuring information retention and recall, can transcend the fundamental differences between spatial and symbolic information. Figures 5 and 6 support this contention. With respect to the retention and recall of symbolic information, the concrete interface metaphor was not observed to be any less effective than the abstract interface metaphor. However, a significant different exists between the two interface types for spatial information retention. Therefore, for the task (i.e., information browsing) used in this particular study, it appears that a concrete interface metaphor can effectively support the retention of both symbolic and spatial data thus extending cognitive fit theory.

The third area that was tested was the polarization of user domain familiarity, which was used to instantiate the relative strength of a user’s mental model. While these results were inconclusive, interesting insights were discovered. Gentner's (1989) work suggests that users with weak domain familiarity are less effective in an abstract domain as compared to users with strong domain familiarity. The results of our study showed no difference between the two domain familiarity groups with respect to the symbolic information. However, there were noticeable differences between these groups when spatial information was presented. Also, the domain familiarity differences were more apparent after the 2-day lag. As stated earlier, future studies will need to be conducted to arrive at a conclusive result.

6.1.2 Practical Implications: From a pragmatic perspective, this study presents results that have the potential to be extremely useful. With the emergence of electronic commerce as a viable means to conduct business, the need to design effective interfaces that provide organizations with a competitive advantage is imperative. This study provides some valuable insights as to how to accomplish this task.

Results indicated that a concrete interface metaphor caused users to retain more information about the organization, particularly any domain specific information that is presented in graphical form. Therefore, if an organization's user base frequently interacts with a physical domain (i.e., business, customer, or product domain), this type of interface will be more effective than the de facto frame interface metaphor. Hence, the decision to utilize this type of interface depends on the nature of the business domain.

The type of information that the organization wishes to present is another issue. If it is symbolic, the results reveal no significant different between the two types of interface metaphors. Conversely, spatial information may be better served by a concrete interface metaphor. If the organization wishes to present a combination of symbolic and spatial information, which is often the case, the practical solution may be to combine the two interface methods. Potentially, this could be explored in more detail in future research.

6.2. Limitations

The research presented from this study contributes both theoretically and pragmatically to the field of Information Systems. However, it is has its limitations. First, the fact that the study was conducted in a laboratory setting has both positive and negative implications. On the positive side, the heavy emphasis on control within a laboratory study allows potentially confounding variables to be neutralized. However, this environment can be perceived to be unrealistic, especially from a pragmatic perspective. For instance, an interface designed for an electronic commerce environment will typically be used by an extremely heterogeneous group of users.

Another limitation of this experimental study is the domain of applicability. Specifically, the number of organizations that will find this type of interface design useful may be limited. As stated earlier, one of the primary means for identifying an effective B2C interface metaphor comes from analyzing the physical domain where users must interact as part of the business process (e.g., a movie theater). A business that does not have a physical domain (e.g., Yahoo) may not find the concept of an interface metaphor based on attributes from the business domain as advantageous. However, these organizations can take advantage of other business-related domains such as product attributes (e.g., graphical representation of a product(s)) or the location where customers consume products/services (e.g., home).

7. Recommendations for Future Research

The results from this study provide an excellent foundation for future research in the area of B2C interface metaphors. This study focused on the effect of a concrete interface metaphor on information retention and retrieval. There is an opportunity to explore the effectiveness of this type of interface metaphor in a transaction-based system. Potential dependent variables could be productivity, accuracy, and satisfaction.

Also, the relationship between mental model and cognitive fit theories should be closely examined. The results of this study suggest that a concrete interface metaphor provides a user with the ability to retain both spatial and symbolic information equally well. It has been postulated that the mental model effect compensated for the opposite effect suggested by cognitive fit theory. More studies using different problem domains with both spatial and symbolic task types should be designed in an attempt to isolate this effect.

While the aforementioned issues are not exhaustive, this study identified a number of attractive research opportunities. By exploring these and other related research questions, a deeper understanding of interface metaphors and their relationship to B2C interaction can be attained. With the emergence of electronic commerce as a viable means to conduct business, the opportunity to study B2C interface metaphors appears to be very promising.

8. References

Agarwal, R. & Venkatesh, V. (2002). Assessing a firm’s web presence: a heuristic evaluation procedure for the measurement of usability. Information Systems Research, 13(2), 168-186.

Adleson, B. (1984). When novices surpass experts: the difficulty of a task may increase with expertise. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 483-495.

Ashcraft, M.H. (1994). Human Memory and Cognition. New York: HarperCollins College Publishers.

Baecker, R.M., Grudin, J., Buxton, W., & Greenberg, S. (1995). Readings in human-computer interaction: toward the year 2000. San Francisco: Morgan Kaufmann Publishers, Inc.

Bettman, J.R. (1979). Memory factors in consumer choice: a review. Journal of Marketing, 43(2), 37-53.

Bettman, J.R. & Kakkar, P. (1977). Effects of information presentation format on consumer information acquisition strategies. Journal of Consumer Research, 3(4), 233-240.

Beyer, H. & Holtzblatt, K. (1998). Contextual design: defining customer-centered systems. San Francisco: Morgan Kaufmann Publishers, Inc.

Carroll, J. & Mack, R. (1985). "Metaphors, computing systems, and active learning. International Journal of Man-Machine Studies, 22, 39-57.

Carroll, J. & Thomas, J. (1982). Metaphors and the cognitive representation of computing systems. IEEE Transactions on Systems Man and Cybernetics, 12(2), 107-116.

Carroll, J.M., Mack, R.L., & Kellogg, W.A. (1988). Interface metaphors and user interface design. In Handbook of Human-Computer Interaction, M. Helander (ed.), North-Holland: Elsevier Science Publishers B.V., 67-85.

Chase, W.G. & Simon, H.A. (1973). Perception in chess. Cognitive Psychology, 4, 55-81.

Chau, P.Y.K., Au, G., & Tam, K.Y. (2000). Impact of information presentation modes on online shopping: an empirical evaluation of a broadband interactive shopping service. Journal of Organizational Computing and Electronic Commerce, 10, 1-22.

Chi, M.T.H, Feltovich, P.J., & Glaser, R. (1981). Categorization and representation of physics problems by experts and novices. Cognitive Science, 5, 121-152.

Coyne, R. (1995). Designing information technology in the postmodern age: from method to metaphor. Cambridge, Massachusetts: The MIT Press.

Dennis, A.R. & Carte, T.A. (1998). Using geographical information systems for decision-making: extending cognitive fit theory to map-based presentations. Information Systems Research, 9(2), 194-203.

Dieberger, A. & Frank, A.U. (1998). A city metaphor to support navigation in complex information spaces. Journal of Visual Languages and Computing, 9(6), 597-622.

Ebbinghaus, H. (1885/1913). Memory: a contribution to experimental psychology. H.A. Ruger and C.E. Bussenius (trans.) New York: Columbia University.

Erickson, T.D. (1990). Working with interface metaphors. In The Art of Human-Computer Interface Design, B. Laurel (ed.), Reading, Massachusetts: Addison-Wesley 65-73.

Gentner, D.R. (1988). "Metaphors as structure mapping: The relational shift. Child Development, 59, 47-59.

Gentner, D. (1989). "The mechanisms of analogical learning. In Similarity and Analogical Reasoning, S. Vosniadou and A. Ortony (eds.), Cambridge, Massachusetts: Cambridge University Press 199-233.

Gerlach, J.H. & Kuo, F. (1991). Understanding human-computer interaction for information systems design. MIS Quarterly, 15(4), 526-549.

Gillan, D.J., Fogas, B.S., Aberasturi, S., & Richards, S. (1997). Cognitive ability and computing experience influence interpretation of computer metaphors. In Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting, 243-247.

Golovchinsky, G. & Chignell, M.H. (1997). The newspaper as an information exploration metaphor. Information Processing and Management, 33(5), 663-683.

Huck S.W. & Cormier, W.H. (1996). Reading statistics and research. New York: Harper Collins.

Jarvenpaa, S.L. and Todd, P.A. (1996). Consumer reactions to electronic shopping on the world wide web. International Journal of Electronic Commerce, 1(2), 59-88.

Jenkins, J.G. & Dallenbach, K.M. (1924). Obliviscence during sleep and waking. American Journal of Psychology, 35, 605-612.

Johnson, E.J & Russo, J.E. (1984). Product familiarity and learning new information. Journal of Consumer Research, 11(1), 542-551.

Johnson, G.J. (1994). "Of metaphor and the difficulty of computer discourse. Communications of the ACM, 37(12), 97-102.

Kay, A. (1990). User interface: A personal view, In The Art of Human-Computer Interface Design, B. Laurel (ed.), Reading, Massachusetts: Addison-Wesley, 191-207.

Kendall, J.E. & Kendall, K.E. (1993). Metaphors and methodologies: Living beyond the systems machine. MIS Quarterly, 17(2), 149-171.

Kim, J. (1999). An empirical study of navigation aids in customer interfaces. Behaviour and Information Technology, 18(3), 213-224.

Koufaris, M. (2002). Applying the technology acceptance model and flow theory to online consumer behavior. Information Systems Research, 13(2), 205-223.

Lackoff, G. & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press.

Lohse, G.L. & Spiller, P. (1998). Electronic shopping. Communications of the ACM, 41(7), 81-88.

Madsen, K.H. (1994). A guide to metaphorical design. Communications of the ACM, 37(12), 57-62.

Nielsen, J. (2000). Designing web usability. Indianapolis: New Riders

Norman, D.A. & Draper, S. (1986). User centered system design. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Inc.

Norman, D.A. (1990). The design of everyday things, New York: Doubleday.

Palmer, J.W. (2002). Web site usability, design, and performance metrics. Information Systems Research, 13(2), 151-167.

Pitkow, J.E. and Kehoe, C.M. (1996). Emerging trends in the WWW user population. Communications of the ACM 39(6), 106-108.

Raskin, J. (1997). Looking for a humane interface: will computers ever become easy to use?. Communications of the ACM, 40(2) 98-101.

Rasmussen, J. (1990). Mental models and the control of action in complex environments. In Mental Models and Human-Computer Interaction 1, D. Ackermenn and M.J. Tauber (eds.), North-Holland, Amsterdam, 41-69.

Rumelhart, D. & Norman, D.A. (1978). Accretion, tuning, and restructuring: three modes of learning. In Semantic Factors in Cognition, J. Cotton and R. Klatzky (eds.), Hillsdale, New Jersey: Lawrence Erlbaum Associates, Inc. 37-53.

Rumelhart, D. (1980). Schemata: the building blocks of cognition. In Theoretical Issues in Reading Comprehension, R. Spira, B. Bruce, and W. Brewer (eds.), Hillsdale, New Jersey: Lawrence Erlbaum Associates, Inc. 33-58.

Satzinger J.W. & Olfman, L. (1998). User interface consistency across end-user applications: the effects on mental models. Journal of Management Information Systems, 14(4), 167-193.

Shaft, T.M. & Vessey, I. (1998). The relevance of application domain knowledge: characterizing the computer program comprehension process. Journal of Management Information Systems, 15(1), 51-78.

Shayo, C. & Olfman, L. (2000). The role of training in preparing end users to learn related software. Journal of End-User Computing, 12(1), 3-14.

Smith-Jentsch, K.A., Campbell, G.E., Milanovich, D.M., & Reynolds, A.M. (2001). Measuring teamwork mental models to support training needs assessment, development, and evaluation: two empirical studies. Journal of Organizational Behavior, 22, 179-194.

Swink, M. & Speier, C. (1999). Presenting geographic information: effects of data aggregation, dispersion, and users’ spatial orientation. Decision Sciences, 30(1), 169-195.

Turkle, S. (1995). Life on the Screen, London: Simon and Schuster.

Umanath, N.S. & Scammell, R.W. (1988). An experimental evaluation of the impact of data display format on recall performance. Communications of the ACM, 31(5), 562-570.

Umanath, N.S., Scammell, R.W., & Das, S.R. (1990). An examination of two screen/report design variables in an information recall context. Decision Sciences, 21(1), 216-240.

Umanath, N.S. & Vessey, I. (1994). Multiattribute data presentation and human judgment: a cognitive fit perspective. Decision Sciences, 25(5,6), 795-824.

Vaughan, M.W. (1998). Testing the boundaries of two user-centered design principles: metaphors and memory load. International Journal of Human-Computer Interaction, 10(3), 265-282.

Vessey, I. (1991). Cognitive fit: a theory-based analysis of the graphs versus tables literature. Decision Sciences, 22(2), 219-241.

Vessey, I. & Galletta, D. (1991). Cognitive fit: an empirical study of information acquisition. Information Systems Research, 2(1), 63-84.

Vessey, I. & Glass, R. (1998). Strong vs. weak: approaches to systems development. Communications of the ACM, 41(4), 99-113.

Vora, P. (1998). Human factors methodology for designing web sites. In Human Factors and Web Development, C. Forsythe, E. Grose, and J. Ratner (eds.), Mahwah, New Jersey: Lawrence Erlbaum Associates, Inc., 153-172.

Washburne, J.N. (1927). An experimental study of various graphic, tabular and textural methods of presenting quantitative material. Journal of Educational Psychology, 18(6), 361-376.

Zinkhan, G.M. & Watson, R.T. (1998). Electronic commerce: a marriage between management information systems and marketing. Journal of Market Focused Management, 3, 5-22.

Figure 1: Aspects of Metaphors (Adapted from Baecker, et al., 1995)

Figure 3: Research Model Highlighting Main Effects

Figure 4: Research Model Highlighting Simple Main Effects

Table 7. Summary of hypothesis testing

| |Information Retention |

|Research Hypothesis |Same Day |2-Day Lag |

|#1 An interface metaphor that is based on concrete business domain attributes will enable a user to |( |( |

|retain significantly more information when compared to an interface metaphor that is based on abstract | | |

|business domain attributes. | | |

|#2: A user who possesses strong familiarity of the business domain will retain significantly more | |( |

|information than a user who possesses weak familiarity of the business domain | | |

|#3a: When using an interface metaphor based on abstract business domain attributes, the amount of | | |

|symbolic information retained by SDFs will be significantly MORE than the amount of symbolic | | |

|information retained by WDFs. | | |

|#3b: The amount of symbolic information retained by WDFs when using an interface metaphor based on | | |

|concrete business domain attributes will be significantly MORE than the amount of symbolic information | | |

|retained by WDFs when using an interface metaphor based on abstract business domain attributes. | | |

|#4a: When using an interface metaphor based on abstract business domain attributes, the amount of | | |

|spatial information retained by SDFs will be significantly MORE than the amount of spatial information | |( |

|retained by WDFs. | | |

|#4b: The amount of spatial information retained by WDFs when using an interface metaphor based on | | |

|concrete business domain attributes will be significantly MORE than the amount of spatial information |( |( |

|retained by WDFs when using an interface metaphor based on abstract business domain attributes. | | |

|#4c: The amount of spatial information retained by SDFs when using an interface metaphor based on | | |

|concrete business domain attributes will be significantly MORE than the amount of spatial information |( |( |

|retained by SDFs when using an interface metaphor based on abstract business domain attributes. | | |

( - Significant at ( = .05

( - p = .062

Figure 5. Profile plot of symbolic information retention (2-day re-test)

Figure 6: Profile plot of spatial information retention (2-day re-test)

8. Appendix

Exhibit 1(a): Sample Screen from the Concrete Interface Metaphor

Exhibit 1(b): Sample Screen from the Abstract Interface Metaphor

-----------------------

[1] NOTE: Within this paper, the terms "user", "consumer", and "customer" are considered to be synonymous.

[2] Because this study included two different information retention instruments, the results from the instrument that was administered after a 2-day lag will be referred to as the “re-test” results.

[3] Two separate ANOVAs were run for the symbolic and spatial retention data sets, respectively.

-----------------------

[pic]

Figure 2. Aspects of mental models

(Norman, 1986)

Table 1. ANOVA results for overall information retention scores

|Source of Variation |Sum of Squares |DF |Mean Square |F-Ratio |Prob. > F |

| | | | | | |

|Domain Familiarity |20.04545 |1 |20.04545 |1.4091 |.2386 |

|Mode of Interface |418.91909 |1 |418.91909 |29.4463 |.0001 |

|Domain Familiarity*Interface |48.01136 |1 |48.01136 |2.8756 |.0936 |

| | | | | | |

|Error |1195.000 |84 |14.226 | | |

|C-Total |1674.8636 |87 | | | |

Table 3. Average information retention scores

(Averages in parentheses denote results from the 2-day re-test)

[pic]

Table 5. Multiple mean comparisons for symbolic information retention

|Group |SC |SA |WC |WA |

| |Same-Day |2-Day |Same-Day |2-Day |Same-Day |2-Day |Same-Day |2-Day |

|SC |-------- |-------- | |

|SA |.846 |.711 |-------- |-------- | |

|WC |.682 |.977 |.991 |.913 |-------- |-------- | |

|WA |1.000 |.074 |.846 |.506 |.682 |.178 |-------- |-------- |

Table 4. Means across treatment groups

| |Symbolic |Spatial |

| |Same-Day |2-Day |Same-Day |2-Day |

|SC |6.86 |5.22 |10.23 |7.09 |

|SA |7.36 |4.64 |6.73 |4.14 |

|WC |7.54 |5.00 |9.95 |6.41 |

|WA |6.86 |3.86 |4.91 |2.09 |

[pic]

[pic]

[pic]

Table 2. ANOVA results for overall information retention scores – Re-Test

|Source of Variation |Sum of Squares |DF |Mean Square |F-Ratio |Prob. > F |

| | | | | | |

|Domain Familiarity |76.40909 |1 |76040909 |5.4909 |.0215 |

|Mode of Interface |445.5000 |1 |445.5000 |32.0145 |.0001 |

|Domain Familiarity*Interface |20.04545 |1 |20.04545 |1.4405 |.2334 |

| | | | | | |

|Error |1168.9091 |84 |13.916 | | |

|C-Total |1710.8636 |87 | | | |

[pic]

[pic]

Table 6. Multiple mean comparisons for spatial information retention

|Group |SC |SA |WC |WA |

| |Same-Day |2-Day |Same-Day |2-Day |Same-Day |2-Day |Same-Day |2-Day |

|SC |-------- |-------- | |

|SA |.001 |.002 |-------- |-------- | |

|WC |.990 |.833 |.003 |.030 |-------- |-------- | |

|WA |.000 |.000 |.187 |.062 |.000 |.000 |-------- |-------- |

[pic]

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download