Conceptual Background and Research Model

---

TAM (Davis, 1989) and its TRA theory base (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975) are central theories in the adoption of new IT. TAM and its relationship to TRA have been discussed extensively by previous research [e.g., Karahanna et al. (1999); Davis et al. (1989)] and will not be elaborated on in detail here. Suffice it to say that research has shown the applicability of TAM that PU and PEOU are antecedents of use-intentions to a wide variety of IT, to both experienced and novice users (Karahanna et al., 1999), and across expertise levels (Taylor & Todd, 1995). At the basis of TRA and TAM is the assumption that reasoned action is the motive behind behavioral intentions. TRA posited two sets of antecedents affecting behavioral intentions: attitude based on beliefs about behavioral outcomes and subjective norms that are influenced by the normative expectations of other people. When Davis (1989) and Davis et al. (1989, 1989) adapted TRA to IT adoption in TAM, they focused on two behavioral beliefs, PU and PEOU. Since these two beliefs directly affect IT use-intentions, and since social norms were insignificant in the case of IT adoption, the more parsimonious model of TAM (Davis, 1989) contains just these two beliefs as the antecedents of IT acceptance without the mitigating effect of attitude that was suggested by TRA and without social norms. Later research showed that social norms might be important only prior to the actual use of an IT (Karahanna et al., 1999).

Subsequent research has mostly adopted this parsimonious model of TAM with just PU and PEOU as predictors of IT use or intended IT use (e.g., Adams et al., 1992; Subramanian, 1994; Hendrickson et al., 1993; Igbaria et al., 1995; Gefen & Straub, 1997; Agarwal & Prasad, 1999; Gefen & Keil, 1998), showing that it applies to a wide range of IT, including e-commerce adoption (Gefen & Straub, 2000, 2003). Based on that previous research, it is assumed that TAM applies also to the continued use of an IT:

H₁: The PU of an IT will increase users' intentions to continue using it.

H₂: The PEOU of an IT will increase users' intentions to continue using it.

H₃: The PEOU of an IT will increase users' PU of it.

An alternative reason behind behavioral intentions-and the objective of this study-is that people continue to do what they are habitually used to doing, without devoting much thought or rational analysis to behaviors they practice out of habit. In other words, all else being equal, past behavior should be an indicator of future behavior in the case of habitual behavior. The theoretical justification for the effect of habit is the "null hypothesis," stating that, all else being equal, it is to be expected that there will be no significant difference between the past and the present. Specifically, it is assumed that habitual previous behavior in a given context will predict behavioral intentions in the same context. Research indeed shows that habitual behavior leads to continued behavior of the same type (Aarts, Verplanken & van-Knippenberg, 1998). Such an effect is a central element in a Triandis' (1971) theory of attitude and attitude change. According to Triandis (1971), behavioral intentions are the product of attitude, social norms, and affect caused by habit. Attitude and social norms correspond to the same named constructs in TRA, and have been operationalized as such by previous MIS research (e.g., Thompson et al., 1994). However, habit, the third element in this theory has not been examined in the context of IT adoption. Habit affects behavioral intentions by the favorable or otherwise feelings (affect) a person has toward an activity based on previous habitual activities. Habit, according to Triandis (1971), influences behavioral intentions over and above attitude and social norms by creating such a feeling toward a behavior.

Another theoretical reason why habit should influence behavior is that once a behavior becomes a habit, then being a well-practiced behavior it becomes automatic and as such is done without conscious decision (Ouellette & Wood, 1998). And, thus, the intention to continue with habitual behavior is an automated decision that is not preceded by a cognitive evaluation process (Aarts et al., 1998). Indeed, when habit is strong, people rely much more on their habit than they do on external information and on choice strategies. Moreover, even when attention to the choice process is increased through manipulation, it does not override the effect of habit (Verplanken, Aarts & Van-Knippenberg, 1997). When habit is strong, people may actually choose to select what information they base their decisions upon (Aarts et al., 1998), biasing their behavioral beliefs toward their own habits. Habit may be a significant predictor of future behavior also because of the human learning process by means of which an assessment of relevant issues is considered early on when a novel situation is encountered but is replaced by habit once the novel behavior becomes routine (Shekhter & Potapova, 1991). Extending this idea to IT adoption implies that when deciding whether to accept a new IT, PU and PEOU should be significant predictors of its usage, because these important aspects of an IT should contribute to the decision process. But, once the use of a specific IT becomes routine, habit should become a primary predictor of its use, because routine activities, such as the continued habitual use of a specific IT, are more influenced by habit.

Indeed, research comparing TRA and related theories with habit as an antecedent of behavioral intentions has shown that habit directly affects behavioral intentions way and beyond attitude and social norms (Leone, Perugini & Ercolani, 1999; Trafimow, 2000; Trafimow & Borrie, 1999; Verplanken, Aarts, Knippenberg & Moonen, 1998); habit will even interfere with the adoption of new behavior (Quinn, Spreitzer & Brown, 2000) and will increase the continuation of existing behavior (Campbell & Cochrane, 1999; Ouellette & Wood, 1998).

Applying these findings to IT adoption suggests that users' behavioral intention to continue using a specific IT, in this case a given website they have habitually used in the past, will be increased directly due to their habit of going to that specific website. In practical terms, this hypothesis states that a person who habitually uses a certain website, or any other software package for that matter, will intend to continue using it simply out of being used to doing so. This will be done regardless of any perceived behavioral outcomes relating to the IT.

H₄: Habitual previous activity with an IT will increase users' intentions to continue using it.

Through habitual prior use of an IT and the knowledge that is gained by doing so, users should learn more about the IT, including how to operate it and how to gain more advantage out of it. This increased understanding of the IT and how to use it should be reflected in increased user perceptions that the IT is easy to use, because an experienced user will know better how to operate it. Such increased understanding should also result in a greater awareness among the users of its potential usefulness. Indeed, users who are experienced with an IT perceived it as more useful and easier to use than users with only limited experience with it (Karahanna et al., 1999). Accordingly, it is hypothesized that the experience gained through previous habitual use of an IT, in this case a website, should increase user perceptions of its ease of use and usefulness:

H₅: Habitual previous activity with an IT will increase user perception of its usefulness.

H₆: Habitual previous activity with an IT will increase user perception of its ease of use.

---