Shopping at www.libri.de - What keeps me there, what makes me leave?

Contents

I. Introduction and Overview of Thesis

II. Theoretical Analysis
1 Scientific Context of Approach-Avoidance Conflict Model
1.1 Behaviourism embedding Approach-Avoidance Conflict Model
1.2 Reflection of Behaviourism in the Context of Online Shopping
2 Theory of Approach-Avoidance-Conflict
2.1 Miller’s Conflict Model
2.2 Critics and further Development of Miller’s Assumptions
2.3 Approach and Avoidance Motivation in Personality
2.4 Summary of Approach-Avoidance Conflict Model
3 Introduction into Online Shopping
3.1 Online Shopping - a Part of E-Commerce
3.2 Penetration of E-Commerce and Online Shopping
3.3 Typology of Online Shoppers
4 Approach-Avoidance Conflict in Online Shopping Situation
4.1 Perceived Customer Value
4.1.1 Costs saving
4.1.2 Time Saving
4.1.3 Convenience
4.2 User Interface
4.2.1 Davis ’ Technology Acceptance Model
4.2.2 Effective Website Design
4.3 Risk Perception
4.3.1 Product Performance Risk
4.3.2 Financial Risk
4.3.3 Psychological Risk
4.3.4 Time Risk
4.3.5 Social Risk
4.3.6 Physical Risk
4.4 Summary of Approach-Avoidance Conflict when Online Shopping

III.Empirical Analysis
5 Research Question and Hypotheses
5.1 Research Question
5.2 Hypotheses
6 Pre-Test
6.1 Aims of Pre-Test
6.2 Results of Pre-Test
7 Online Simulation
7.1 Method of Simulation
7.1.1 Participants
7.1.2 Instrument
7.1.3 Realisation
7.2 Results of Simulation
7.2.1 Descriptive Data Analysis
7.2.2 Testing of Hypotheses
7.2.3 Explorative Data Analysis
8 Reflection of Method and Discussion of Results
8.1 Reflection of Method
8.2 Discussion of Results

IV. Summary

V. Bibliography

VI. Appendices

Appendix A - Tables. 97 Appendix B - Instrument Pre-Test.106 Appendix C - Instrument Simulation 108 Appendix D - Raw Data

Acknowledgements

To everybody who contributed in one way or another…

Leider lässt sich eine wahrhafte Dankbarkeit mit Worten nicht ausdrücken.

Johann Wolfgang von Goethe (1749 - 1832)

deutscher Dichter der Klassik,

Naturwissenschaftler und Staatsmann

Figures

Figure 1: Approach-avoidance conflict according to Miller

Figure 2: Approach-avoidance conflict according to Maher

Figure 3: Conflict types one to four according to Herkner

Figure 4: Conflict types five to eight according to Herkner

Figure 5: Conflict types according to Yelen

Figure 6: Areas of e-commerce

Figure 7: E-commerce sales in percent of total turnover

Figure 8: Technology acceptance model

Figure 9: Graphical illustration of online simulation procedure

Figure 10: Graphical overview of means of approach tendencies

Figure 11: Graphical overview of means of avoidance tendencies

Figure 12: Cook’s influence statistics of cases

Figure 13: Development of approach and avoidance tendencies per group of queries

Figure 14: Sequence of surfing

Tables

Table 1: Cross tab with setting and purchase simulated

Table 2: Overview of explained variance and factor loadings for each point in time

Table 3: Means and standard deviations of identified factors

Table 4: Means and standard deviations of factors grouped in number of queries

Table 5: Sphericity and variance analysis repeated measures grouped in number of queries

Table 6: Means and standard deviations of interface with four measures

Table 7: Means/medians and standard deviations/ranges of distances of factors

Table 8: Classification table of binary logistic regression

Table 9: Distances in equation of binary logistic regression

Table 10: Expected and observed counts of purchases considering risk disposition

Table 11: Means/medians of onsite condition minus means/medians of online condition

Table 12: Frequency of „missing“

Table 13: Reasons indicated for non-purchase

Table 14: Positive feedback to the simulation

Table 15: Negative feedback to the simulation

Abbreviations

illustration not visible in this excerpt

I. Introduction and Overview of Thesis

Different motivation tendencies, aspirations and fears connected to the same action quite often happen to contradict each other. A person looks at the positive (= approaching) and negative (= avoiding) sides when aiming towards a goal or reaching a decision. For example, on the one hand a student wants to go swimming, as it is a nice warm summer day, but on the other hand the student should be studying for his upcoming exam. The student undergoes a conflict about what to do.

The following work is based on this special kind of conflict called approach-avoidance conflict. In consumer behaviour literature (cf. Solomon, Bamossy, & Askegaard, 2001), this conflict is often mentioned regarding the general decision of whether or not a certain product will be bought. Positive as well as negative aspects of the product are judged and weighted when reaching a decision. Thus, it is assumed that this kind of conflict also occurs in the context of online shopping. Online shopping has both positive as well as negative aspects. It encompasses approaching (where beneficial rewards are expected) and avoiding tendencies (where harmful rewards are expected). In the context of online shopping, the approach tendency could be for example the convenience of shopping from home. It is possible to get a book ordered and delivered without even leaving the house. At the same time, potential online shoppers might worry about credit card fraud if payment is demanded by credit card. The latter would represent the avoidance tendency. Both tendencies are competing against each other. Which one is stronger? The consumer has to make a choice because the person cannot shop online to benefit from the convenience and at the same time not shop online to avoid possible credit card fraud. The two responses are incompatible.

Regarding online shopping, several tendencies both strengthening and weakening determinants occur. The goal is to purchase online. It will not be reached if avoidance tendency is stronger than approach tendency.

Thus, the basic idea of the study is to explore the relevant approach and avoidance factors when online shopping. Besides its identification, it is also of interest how dominant their impact is perceived. First, it is questioned whether changes of this dominance occur during the shopping process. Second, which factors are predicting the purchase decision. Third, whether the purchase decision is influenced by general harm avoidance since risk perception presents the main limiter of online shopping decisions.

In the beginning of the thesis, an overview of the scientific context is given. The definition of the approach-avoidance conflict dates back to the 1940s. Thus, it is considered necessary to present the applicable paradigm in which it is embedded. Then, the theoretical analysis of the approach-avoidance conflict with its modifications, extensions, and limitations is presented. After discussing the theory, an overview of online shopping is given. Approach and avoidance factors influencing the decision to purchase are introduced and discussed.

The second part of the thesis deals with the empirical investigation of the approach- avoidance conflict in the context of online shopping. It is analysed by simulating an online purchase in an online book shop. First of all, research question and hypotheses are stated, followed by the description of the pre-test with its aims and results. Then, the actual online simulation, its method and results is presented, followed by a discussion of results and a reflection of method. Finally, the work is summarised and a general outlook on further investigation is given.

II. Theoretical Analysis

The first part of this work is a theoretical analysis regarding the approach-avoidance conflict. The scientific context and background of the approach-avoidance is discussed in chapter 1. The theory of approach-avoidance conflict is introduced and discussed in chapter 2. After presenting conflict theory, the field of investigation (i.e. online shopping) is described in chapter 3. Finally, the approach-avoidance conflict in online shopping situations, in particular the approach-avoidance tendencies when online shopping, is analysed.

1 Scientific Context of Approach-Avoidance Conflict Model

Since it is also necessary to look behind a theory in order to fully understand it, the following chapter aims to describe the scientific context when the conflict model was postulated. Thus, behaviourism embedding the approach-avoidance conflict model is presented in 1.1. Furthermore, it is questioned if this paradigm is also appropriate for online shopping. In 1.2, a reflection of behaviourism in the context of online shopping is given.

1.1 Behaviourism embedding Approach-Avoidance Conflict Model

Theories are influenced by the respective paradigm. Paradigm stands for general assumptions and laws within a field on which theories are based (cf. Kuhn, 1970, pp. 57). The paradigm embedding the approach-avoidance conflict model is behaviourism. One of the most famous behaviourists is the US-American psychologist Watson (1878-1958). In Watson’s book “Psychology from the standpoint of a behaviourist” (originally published 1919) he describes Psychology as

…a Science of Behavior. - Psychology is that division of natural science which takes human activity and conduct as its subject matter. It attempts to formulate through systematic observation and experimentation the laws and principles which underlie man’s reactions. (Watson, 1994, p. 1)

Behaviourists focus only on a stimulus and its response in order to predict and control behaviour. They leave out everything in between, such as emotions, motivation, and attitudes - this is also referred to as the black-box model or S-R (stimulus-response) model, respectively. Only statements about physical and psychological principles which can be tested objectively are permitted (Vanecek, 1998, p. 13). Learning theories have been based on reinforcement, which was also used in the experiments for the testing of the approachavoidance conflict as described below.

This strong suppression of the organism between stimulus and response has been widely criticised. Therefore, within neo-behaviourism the S-O-R model (stimulus-organism- response model) was introduced. Thus, it does not totally exclude the organism (“O”) between stimulus and response, but it also partially includes unobservable intervening variables (Felser, 2001, p. 14). Approach-avoidance conflict was developed in the thinking of neo-behaviourism. Regarding online shopping, the stimulus would be the online shop and the response would be the purchase with the intervening conflict in between. Whether this behaviouristic model is applicable for online shopping is discussed in the following section 1.2.

1.2 Reflection of Behaviourism in the Context of Online Shopping

The two basic types of models to explain consumer behaviour are the S-R models and the S-O-R models described above (Schneider, 2004, pp. 33). Foscht (1999, p. 145) thinks that this approach is not enough to understand consumer behaviour in the context of new media even if the models are extended. He (Foscht, 1999, p. 150) argues that within those models, human behaviour is always seen as a reaction to stimuli, which is not applicable for new media. In the new media context, the active role of the consumer shall be emphasized. In Foscht’s view (1999, p. 145) the solution to this problem lies in constructivism and system theory, respectively.

Constructivists say we construct the world where we live in. Reality cannot be found, but is invented. Therefore, it is not possible to find absolute verity or even be able to know whether that what one experiences corresponds to a world independent from oneself. The only surety is that the person experiences something. (cf. Glasersfeld, 1985)

Thus, the aim of science should not be to find absolute truth, but to have theories viable in a given environment. The principle of key and lock is an example of this viability. The key needs to open the lock; it is not important whether it is the exact representation as long as it fulfils its purpose. (cf. Foerster, 1993) Because of this self-construction of our reality, there have to be as many realities as there are people or systems, respectively.

Kroeber-Riel and Weinberg (2002, pp. 570) talk about a second reality of the consumers - the media environment. But this is not far enough for Foscht (1999). According to a constructivist understanding, however, there is only one world for every individual. The individual represents a closed and autonomous system, constructing its own reality.

A system’s structure and organisation is created by interactions with the environment. The system is organised by itself (i.e. autopoesis) with system-innate components that require each other (Baitsch, 1993, p. 9). It is possible to have instable sub-processes, but the whole system is held constant. Despite this organisational closeness, the system is energetically open. This permeability is necessary to survive as an organisation (Schmidt, 1994, cited in Foscht, 1999, p. 146). An example of a system could be the market surrounded by its environment, the economic world, the cultural world and the world as a whole. The market covers the subsystems demand and offer. The market players also represent a subsystem e.g. an online shop. The elements, subsystems or systems (including the environment) are connected interdependently. Thus, mutual interaction is essential (cf. Hünerberg, 1984 p. 6).

Interactions are the basis of communication. Because of the changing environment, continuous interaction is necessary for survival. External change also causes modification within the system. However, this causation must not be seen as externally controlled stimuli, but only represents animations or modulations for the acting individual (Foscht, 1999, pp. 146).

Because of this active role of the individual, it is not enough to explain consumer behaviour on the basis of the S-O-R model. Concluding, Foscht admits that currently there are not enough insights to apply his thoughts and asks for further research (Foscht, 1999, p. 150).

Following Foscht’s considerations, the approach-avoidance conflict theory does not seem sufficient for understanding the online shopping process at first sight. Why is it used in the current study then? Contrary to the original studies about the approach-avoidance conflict, approach as well as avoidance tendency is not only represented by one dimension, but a wide range of factors. Additionally, the influence of personality on risk perception is also incorporated. Thus, the application of the approach-avoidance conflict to understand online shopping seems appropriate, in particular, with respect to the fact that Foscht himself sees the limited use of the system theoretical model at present. Further research is required before being able to apply his suggestions.

2 Theory of Approach-Avoidance-Conflict

Deciding which product to buy is often connected with a psychological conflict (Antonides & Raaij, 1998, 258). If the choice has positive as well as negative aspects then it is called approach-avoidance conflict, which is defined as a motivational conflict (Kroeber-Riel & Weinberg, 2002, pp. 161).

Besides the approach-avoidance conflict, there are also other types of motivational conflicts. Approach-approach conflict occurs when deciding between two attractive possibilities Analogously, the decision between two negative alternatives is called avoidance-avoidance conflict (cf. Antonides & Raaij, 1998, pp. 259; Miller, 1944, p. 432 ). The decision-making process between two alternatives - each alternative encompassing both positive as well as negative aspects - is called double approach-avoidance conflict (cf. Antonides & Raaij, 1998, p. 260; Miller, 1944, p. 433).

The current chapter aims to give a comprehensive overview of the approach-avoidance conflict theory. First of all, the original conflict model postulated by Miller is presented in section 2.1, followed by criticisms and further development of the conflict model in section 2.2. After that, it is shown in section 2.3 that approach and avoidance motivation is an underlying concept in personality. The chapter closes with a summary and conclusion of the presented argumentation in section 2.4.

2.1 Miller’s Conflict Model

According to Miller, conflict is created by “ competition between incompatible responses ” (Miller, 1944, p. 431). There are many different types of conflicts. The best known conflict is termed approach-avoidance conflict (Weiner, 1992, p. 86). Miller has conducted fundamental studies on approach-avoidance conflict drawn on field theoretical findings (Lewin, 1963) and stimulus-response learning theory (Hull, 1938).

Based on Lewin’s and Hull’s theories, Miller formulated the following six basic principles:

(A) The tendency to approach a goal is stronger the nearer the subject is to it. This is an application of Hull’s principle of the goal gradient and will be called the Gradient of Approach.
(B) The tendency to go away from a place or object avoided is stronger the nearer the subject is to it. This was an extension of the general idea of the gradient of reinforcement to avoidance learning. It will be called the Gradient of Avoidance.
(C) The strength of avoidance increases more rapidly with nearness than does that of approach. In other words, the gradient of avoidance is steeper than that of approach. […]
(D) The strength of the tendencies to approach or avoid varies directly with the strength of the drive upon which they are based. In other words, an increase in drive raises the height of the entire gradient. […]
(E) Below the asymptote of learning, increasing the number of reinforced trials will increase the strength of the response tendency that is reinforced.
(F) When two incompatible responses are in conflict, the stronger one will occur. (Miller, 1964, pp. 99)

Of all the postulates tied in with existing theories, only principle C was new, and according to Miller (1964, pp. 99) it is necessary to explain the behaviour of going part way and then stopping. Nevertheless, it was the most critical assumption (Epstein & Fenz, 1962; Losco & Epstein, 1977).

To verify these principles, Miller (Miller, 1944; 1964) and Brown (Brown, 1948 cited in Weiner, 1992, p. 91) were the first to test Miller’s approach-avoidance model experimentally. They carried out various experiments with albino rats. First of all, the rats were deprived of food. Then they were trained to run down an alley in order to get food, which was placed at the other end and made distinctive by a light source (appetence training). After this training period, electric shocks were administered when taking food (avoidance training). Thus, positive as well as negative attributes were connected to the goal.

Theoretical Analysis 14

Subsequently, in the test trial, no electric shock was given when reaching food. In this situation, the animals approached the goal, but then stopped at a certain point, which is the crossing point of gradients (= point of intersection), in figure 1. This point represents equal strength of avoidance and approach tendencies. Depending on the number of days of food deprivation, the number of training trials and the strength or number of shocks, the point of intersection varies. The stronger hunger or the weaker shock the nearer to the goal the animal had come.

Figure 1 shows a graphical representation of the approach-avoidance conflict. Distance (i.e. space) is plotted on the abscissa and strength on the ordinate. Strength was measured by applying a harness device to each rat in order to assess how hard it pulled. The linear illustration of gradients is just a matter of simplification.

Figure 1: Approach-avoidance conflict according to Miller (1944, p. 434)

illustration not visible in this excerpt

The abovementioned assumptions were developed from animal experiments. Miller has assumed that the basic mechanisms of learning and conflict between humans and other mammals are similar enough in order to generalise the results to human behaviour. (Miller, 1964, p. 98)

2.2 Critics and further Development of Miller’s Assumptions

Besides the animal experiments described above (Brown, 1948 cited in Weiner, 1992, p. 91; Miller, 1944, 1964), approach-avoidance conflict has also been investigated with humans. On the one hand, it was investigated in real settings such as sports situations like parachuting (cf. Epstein & Fenz, 1962; Schwarzinger, 1980), and rock climbing (cf. Heitzlhofer, 1978) as well as regarding the year 2000 problem (Sturm, 2001) and the examination for the driving licence (Klinger, 1990). On the other hand, approachavoidance conflict was also observed in rather experimental settings with humans such as betting situations (cf. Losco & Epstein, 1977; Yelen, 1979, 1985). Within this thesis, approach-avoidance conflict theory is used to explain consumer behaviour. In particular, the decision-making process of online shopping is investigated.

As can be seen in the examples above, Miller’s approach-avoidance conflict occurs also within humans. Applying Miller’s results one-to-one to human behaviour poses a logical problem for Maher (1964, p. 288). This problem is particularly acute when approach tendency is caused by situational cues and not by internal physiological ones like hunger. He has considered the approach tendencies mainly as learned behaviour tendencies illustrated with Epstein and Fenz’s need for adventure, excitement and prestige (Epstein & Fenz, 1962, p. 288). In the experiments with rats, only avoidance tendency is learned. However, this is not the same in the case of humans, where both avoidance and approach tendency represent learned behaviour. Consequently, when both are equally learned, there is no logical reason that the gradient of approach should be steeper, but both gradients should have a similar slope. Thus, Maher (1964, p. 289) suggested the following additional postulate to those of Miller. „The stronger of two competing response tendencies mediates behaviour only when its absolute momentary strength exceeds that of the weaker by an amount which is a constant fraction of the weaker.“ (Maher, 1964, p. 289) In other words, the stronger one of the approach and avoidance gradients determines behaviour and both take a parallel course. An illustration of this can be seen in figure 2.

Although the gradients are parallel, Maher also applies an equilibrium point corresponding to Miller’s point of intersection (Miller, 1964, p. 101). The locus of the equilibrium point depends on the proportion of the weaker gradient to the total strength of tendency. When it has reached certain strength, oscillatory behaviour occurs.

Figure 2: Approach-avoidance conflict according to Maher (1964, p. 289)

illustration not visible in this excerpt

Other authors (cf. Herkner, 1975, p. 52; Losco & Epstein, 1977, p. 360; Weiner, 1992, p. 87) also emphasise the difference between inner physiological cues and situational ones. For instance, Herkner considers the approach-avoidance conflict applicable for humans but suggests some adaptation of Miller’s model (Herkner, 1975). Just like Maher (1964), for Herkner the avoidance tendency is not necessarily steeper than the approach tendency. Contrary to Maher (1964), Herkner (1975) does not add assumptions, but limits them to only three assumptions offering a more general theory.

1. There is an approach gradient
2. There is an avoidance gradient
3. Behaviour in approach-avoidance-conflicts corresponds to the difference of response strength between approach and avoidance gradients

Based on the three assumptions, Herkner (2001, pp. 94) differentiates eight types of conflicts, illustrated in figure 3 and 4, including also parallel types as Maher (1964) has suggested.

Figure 3 shows four possible scenarios where avoidance is stronger at the goal than approach tendency. Each scenario shows two boxes, the left one illustrates the conflict in the same manner as Miller did (type four represents Miller’s conflict model). The right box shows the performed behaviour i.e. the difference between the two gradients. The dotted line indicates the avoidance gradient and the solid line represents the approach gradient.

Figure 3: Conflict types one to four according to Herkner (2001, p. 94)

illustration not visible in this excerpt

Figure 4 represents nearly the same types as figure 3 but differs with respect to the fact that approach and avoidance gradients are swapped.

Figure 4: Conflict types five to eight according to Herkner (2001, p. 95)

illustration not visible in this excerpt

Type eight shows the pendant to Miller’s approach-avoidance conflict and is consequently called avoidance-approach conflict (mentioned first by Losco and Epstein (1977, p. 361)) because of the stronger avoidance tendency in the beginning.

Similarly to Herkner, Yelen (1979) also extended Miller’s approach-avoidance conflict model. Yelen combined Miller’s conflict model with Miller’s displacement model. Originally, the critical factor in the conflict model was spatial distance (cf. Miller, 1964) and in the displacement model it was stimulus similarity (cf. Miller, 1948). Instead of distinguishing between spatial and similarity distance, he differentiates the two ways the tendencies were learnt.

The first type is called passive avoidance learning (D'Amato, 1970 cited in Yelen, 1979, p. 346). It means that avoidance tendency is learnt by punishment of approach response. Therefore, it is not independent from the approach tendency and Yelen (1979, p. 346) suggests a single gradient model. But as Maher (1964) has stated that regarding human behaviour, approach tendency is also learnt behaviour. Consequently, the single gradient model cannot be applied to the current research.

If conflict occurs due to independently learnt approach and avoidance tendencies (such as in the case of online shopping), Yelen (1979, p. 326) suggests a two-gradient model. The two-gradient model, following the displacement model, suggests that behaviour depends on motivation tendency with the dominant net strength, i.e. the stronger tendency minus the weaker tendency. This suggestion might be in accordance with Herkner’s model, even though Yelen (1979, p. 343) identifies nine, instead of eight, types of conflict, depicted in figure 5. The additional type proposed by Yelen can be seen in the middle row, the second type. In this type, approach and avoidance gradients are equally steep. Thus, no tendency dominates over the other one and consequently it appears that a real decision cannot be made for this type. It seems appropriate for betting situations, where the probability of winning or losing is estimated to be equal, and therefore net strength would be zero.

Figure 5: Conflict types according to Yelen (1979, p. 343)

illustration not visible in this excerpt

Yelen (1979) experimentally verified these types in betting situations whereas Herkner’s (1975) and Maher’s (1964) critics are only theoretically based. Yelen (1985, p. 391) also replicated his results with continuous response measures, i.e. various positions in the conflict were measured instead of only one single choice from a number of alternatives.

To conclude, for the current study Herkner’s model is considered to be overall more adequate for understanding the decision-making processes than Yelen’s propositions. In consumer research, the decision has to be made whether the consumer buys the product. Saying that both products have equal attraction is not sufficient (which would occur in Yelen’s additional type).

Gjesme (1974) has applied the approach-avoidance conflict to achievement motivation. He differentiates between approach-oriented and avoidance-oriented pupils. The approach- oriented pupils are highly motivated to approach success and have low motivation to avoid failure. The avoidance-oriented ones act vice versa. According to Miller’s conflict model, approach-oriented pupils would increase and avoidance-oriented pupils would decrease their level of performance as a distant future goal approaches in time. These assumptions were confirmed using number of problems attempted as a measure. The hypothesis that avoidance-oriented pupils would have a steeper slope of goal gradient for performance than the approach-oriented pupils received no support. These results are in concordance with the introduced critics of Miller’s postulates (Herkner, 1975; Maher, 1964; Yelen, 1979).

2.3 Approach and Avoidance Motivation in Personality

Personality is associated with a person’s “general and consistent forms of conduct” (Eysenck, 2004, p. 2). Elliot and Thrash (2002) have found that approach and avoidance motivations are underlying concepts of personality. “In approach motivation, behaviour is instigated or directed by a positive/desirable event or possibility whereas in avoidance motivation, behaviour is instigated or directed by a negative/undesirable event or possibility.” (Elliot & Harackiewicz, 1996 cited in Elliot & Thrash, 2002, p. 804)

In order to empirically verify the approach and avoidance motivation, Elliot and Thrash (2002, p. 804) analysed the basic structural dimensions in personality. Regarding these basic dimensions, they (Elliot & Thrash, 2002, pp. 804) identified the focus on trait adjectives, on affective dispositions, and on motivational systems. Each basic structural dimension encompasses two opposed elements which correspond to approach or avoidance motivation, respectively.

Regarding the trait adjectives, two salient models are dominant namely the Big Five Model composed of neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness (McCrae & Costa, 1987) and the Big Three Model encompassing neuroticism, extraversion, and psychoticism (Eysenck & Eysenck, 1985). Indeed, in both models the dimensions Extraversion and Neuroticism are part of it. Extraversion represents characteristics such as sociable, active, and optimistic whereas Neuroticism stands for worry-prone, emotionally unstable, and insecure. In Elliot and Thrash’s (2002, pp. 804) empirical verification of the approach and avoidance motivation, trait adjectives were assessed with Costa and McCrae’s (1992) NEO-FFI.

With respect to the affective dispositions, the dimensions positive and negative emotionality were identified. Positive emotionality is defined as a “broad tendency to experience positive emotion and to engage life in a positive manner” (Elliot & Thrash, 2002, p. 805) whereas negative emotionality represents “a broad tendency to experience negative emotion and to engage life in a negative manner” (Elliot & Thrash, 2002, p. 805). Elliot and Thrash (2002, p. 805) assessed the affective dispositions with Watson and Clark’s (1993) General Temperament Survey (GTS).

Finally, the third approach is represented by motivational systems. Based on Gray (1970), Elliot and Thrash (Elliot & Thrash, 2002, p. 805) identified the behavioural activation system (BAS), which facilitates behaviour and produces positive affect and its opposite, the behavioural inhibition system (BIS), which inhibits behaviour and produces negative affect. The two systems were assessed by Elliot and Thrash (2002, p. 807) with Carver and White’s (1994) BAS and BIS Scales.

Each of the above described approaches (neuroticism and extraversion, positive and negative emotionality as well as BIS and BAS) includes two opposed dimensions. By factor analysis over the six dimensions, it was confirmed that there are two underlying constructs which were labelled approach and avoidance temperament. Approach temperament is shown by the dimensions extraversion, behavioral activation system (BAS), and positive emotionality whereas avoidance temperament can be observed by means of neuroticism, behavioral inhibition system (BIS), and negative emotionality (Elliot & Thrash, 2002, p. 813).

Concluding, it was shown that there is a general neurobiological sensitivity to positive or negative stimuli. In other words, there is a general tendency in personality to approach or avoid a goal which influences behaviour. Consequently, human behaviour does not only depend on the respective situation and/or the stimuli, but also on personality. This seems to underline Foscht’s argumentation (Foscht, 1999) outlined in 1.2 which states that inducing stimuli and analysing its response is not enough to understand consumer behaviour and emphasises the active role of the individual.

2.4 Summary of Approach-Avoidance Conflict Model

To sum up, the authors (Epstein & Fenz, 1962; Herkner, 1975; Losco & Epstein, 1977; Miller, 1944; Yelen, 1979) agree on the existence of an approach-avoidance conflict. When approaching a goal, there is an avoidance tendency as well as an approach tendency competing with each other. Furthermore, they assume that tendencies increase with goal nearness. This increase in strength is also referred to as the “goal looms larger” effect (cf. Förster, Higgins, & Idson, 1998). Contradictory results (Klinger, 1990; Sturm, 2001) also exist, stating that there might also be some additional types of conflict with declining tendency nearer to the goal. But these results are very limited.

In general one can conclude that the resolution of approach-avoidance conflict can be seen in a reduction of Miller’s postulates as well as an increase in the number of types of conflict. There are the two competing approach and avoidance tendencies, which most probably increase with goal nearness. Behaviour occurs according to the tendency with the dominant net strength.

Finally, the approach-avoidance conflict emerges in various situations, provided that the goal aimed at has both positive as well as negative aspects. The chosen field of investigation within this study is online shopping.

3 Introduction into Online Shopping

A general overview about online shopping is given in the current chapter. First of all, the meaning of the term online shopping as well as the borderline to related terms is presented in section 3.1. Then, an illustration of the increasing importance of online shopping is given in section 3.2. Closing, a description of online shoppers is given in section 3.3.

3.1 Online Shopping - a Part of E-Commerce

Often, the term online shopping is used synonymously for e-commerce (Kotler & Bliemel, 2001, p. 1220), but there are differences. Whereas online shopping stands for the distribution of goods via Internet for private use (cf. Borgwardt, 1999, p. 24), e-commerce is the more global term.

Regarding e-commerce, there is a whole range of definitions. A rather general one is the following one from the European Commission:

Electronic Commerce is about doing business electronically. (…) It involves both products (e.g. consumer goods, specialised medical equipment) and services (e.g. information services, financial and legal services); traditional activities (e.g. healthcare, education) and new activities (e.g. virtual malls) (1997, Hermanns & Sauter, 1999, p. 15)

Hence, e-commerce is about electronic transactions involving products, services, and also activities. In addition to that, there is a classification depending who interacts electronically. An overview of interaction partners is shown in figure 6. There are four major ones such as B2C (business-to-consumer), B2B (business-to-business), C2C (consumers-to-consumers), and C2B (consumers-to-businesses). The remaining groups are the relations of businesses and consumers to government (or administration in other words).

Figure 6: Areas of e-commerce (cf. Hermanns & Sauter, 1999, p. 23)

illustration not visible in this excerpt

The business-to-consumer (B2C) segment is defined by online shopping. The most frequent online consumer purchases are books/magazines, music, and computer software (ComCult, 2002, pp. 11; Kotler, 2003, p. 40; Krause, 1999, p. 103). These products are rather standardised and their quality can be easily evaluated.

There are two types of companies selling via the Internet - the “pure-click” companies and the “brick-and-click” companies. Pure-click companies are those that launched a website without any previous existence as a firm and business is only made via the website, whereas brick-and-click companies are pre-existing businesses which have added an online site for information and/or e-commerce (Kotler, 2003, p. 45)

The business-to-business (B2B) segment represents a ten to fifteen times greater segment than B2C. It encompasses auction sites, spot exchanges, online product catalogues, barter sites and other online resources to obtain better prices. (Kotler, 2003, p. 42)

Consumer-to-consumer (C2C) “means that online visitors increasingly create product information, not just consume it.” (Kotler, 2003, p. 43) In chat rooms and Internet interest groups, consumers are able to exchange their opinions about products and companies. Additionally, it covers online auctions such as e-bay (Hermanns & Sauter, 1999, p. 23). The consumers-to-businesses (C2B) segment is also based on the actions of consumers.

Consumers address enterprises, for instance at an online job exchange where job seekers put in ads. (Hermanns & Sauter, 1999, p. 23)

The remaining four groups are the relations of businesses and consumers to administration. Administration-to-Business covers public institutions’ purchases via the Internet. Consumer-to-administration encompasses tax handling of private persons whereas business-to-administration encompasses tax handling of enterprises. The last segment administration-to-administration includes all transactions between public institutions within their home country as well as abroad. (Hermanns & Sauter, 1999, p. 23)

3.2 Penetration of E-Commerce and Online Shopping

Having begun in the United States, the trend of e-business is spreading to other countries. Despite the disappointment after the crash of the dot-coms in 2000 and 2001, enterprises are now investing in e-business again. Investments in information technology generally decreased whereas they rose in e-business budgets. (UNCTAD, 2003, p. 1)

The two most important segments in e-commerce are B2B or B2C, respectively. In particular, business-to-business e-commerce shows the biggest share of sales figures. In 2001, B2B online sales in the United States amounted to $ 995 billion that is 93.3 per cent of all e-commerce in that country. In comparison with US total sales and revenues in 2001, this $ 995 billion represents 14.9 per cent. (UNCTAD, 2003, pp. 20)

The United States takes the leading role regarding e-commerce. But the other countries are converging. In 2001, online trade in the European Union was still less than 1 per cent of all B2B trade whereas it has been forecasted to reach nearly 10 per cent by the end of 2004. B2B e-commerce in Central and Eastern Europe (CEE) is a reality in just three countries, the Czech Republic, Hungary and Poland, which are also members of the European Union by now. In the other CEE-countries, there is nearly no e-commerce at all.

Furthermore, impressive growth rates are also projected in the Asia-Pacific region. It is forecasted from about $120 in 2002 to around $200 billion in 2003 and $300 billion by 2004. B2B e-commerce in the Asia-Pacific region mostly takes place in Japan, Republic of Korea, China and India. In Latin America, B2B e-commerce is mainly driven by Brazil,

Argentina and Mexico. Similarly, e-commerce sales in the African region are mainly carried by South Africa with 80 to 85 per cent. (UNCTAD, 2003, pp. 20)

Parallel to B2B e-commerce, the growth of sales figures in the B2C e-commerce segment is projected. In 2002, online retail sales already amounted to 1.34 per cent of total retail sales in the United States and 1.6 per cent in the European Union. Eastern Europe is expected to retain only a very small share of retail sales. (UNCTAD, 2003, pp. 18) Regarding Western European, a brief overview about total e-commerce sales in 2003 is presented in figure 7.

Figure 7: E-commerce sales in percent of total turnover (Eurostat, 2004, p. 190)

illustration not visible in this excerpt

In general, the northern European countries seem to be more progressive than the southern European countries regarding electronic commerce. Austria is below most other European countries with .9 per cent of total turnover resulting from e-commerce whereas Ireland impresses with 10 per cent. (Eurostat, 2004, p. 190)

10 per cent of global B2C online sales are generated by the Asia-Pacific region, in particular by Japan, Australia, and the Republic of Korea. Equally to the B2B segment, the largest market for B2C e-commerce in Latin America represents Brazil, Argentina, and Mexico. (UNCTAD, 2003, pp. 18)

To conclude, the estimations of these figures vary tremendously. For instance, in 2002 B2C e-commerce in the United States was estimated by Jupiter Research (Feb. 2003) to be $ 40.4 billion and by Forrester (May 2003) to be $ 76 billion. (UNCTAD, 2003, p. 18) Thus, this high gap between the two estimations relativises the accuracy of the figures indicated.

Having Internet access is a prerequisite for e-commerce. Once limited to a certain group, Internet access becomes more and more common. In Austria, 42.7 per cent of all people over 16 used the Internet between March 2002 and 2003, younger people more than older ones. Moreover, 36.2 per cent of all Austrian households had Internet access in 2003 (Statistik Austria, 2004, p. 151). These figures have shown a continuous growth within the last years (Integral, 2004). Burgenland and Carinthia are the federal states with the lowest Internet penetration in Austria - namely 31.6 per cent and 30.1 per cent, respectively (Statistik Austria, 2004, p. 151).

When comparing global Internet penetration rates among countries (worldwide 9.72 per cent), the difference between the developed countries (32.62 per cent) and the developing countries (3.91 per cent) is salient. The leading regions in Internet penetration rates are North America (53.22 per cent), Oceania (33.30 per cent), and Europe (20.79 per cent). (UNCTAD, 2003, p. 5)

3.3 Typology of Online Shoppers

According to Statistik Austria (Statistik Austria, 2004, p. 152), 23 per cent of female and 27.6 per cent of male Austrian Internet users have bought at least one product online for private purposes in 2003. This dominance of male shoppers is shown analogously for the period of January to March 2003 when 15.4 per cent of female and 20.3 percent of male Austrian Internet users purchased goods via the Internet (Statistik Austria, 2004, p. 152).

The majority of Austria’s online shoppers are between 25 and 45 years old (Statistik Austria, 2004, p. 152). The higher the education, the higher is the percentage of online shoppers among Austrian Internet users (Statistik Austria, 2004, p. 153). In general, online shoppers have been identified as being younger, wealthier, better educated and more computer literate than non-Internet shoppers (Swinyard & Smith, 2003, p. 573). These demographics correspond with ComCult’s study (2003, p. 9) about the attractive target group high online consumer. Nonetheless, characteristics of online shoppers are found rather in lifestyle and not in demographics (Bhatnagar & Ghose, 2004, p. 758). Bellman, Lohse and Johnson (1999, p. 37) suggest to look for a “wired” lifestyle such as number of months on the Internet, hours online per week, or searching for product information as well as time starvation in order to identify online shoppers.

Other typologies of online shoppers are based upon primary motives for shopping (Klietmann, 2001; Rohm & Swaminathan, 2004), purchase behaviour (Bhatnagar & Ghose, 2004), or an integration of both and also including demography (Keng, Tang, & Ghose, 2003). Keng, Tang and Ghose (2003) identified six types of online shoppers in Asia from a sample of 3712 usable responses:

- Information surfers

The information surfers love banner ads and click on them often. They look out for promotional offers and have good navigation expertise and online experience. Moreover, they are more likely to be married.

- Comparison shoppers

Comparison shoppers are those who compare product features, prices, and brands before making purchase decisions. Furthermore, they actively look for promotional offers and are between 25 and 29 years.

- On-off shoppers

On-off shoppers are likely to surf the Internet and collect online information but prefer to shop offline. They are experienced in surfing, use bookmarks and the same search engine on a regular basis and often look out for best deals. They tend to be single and in the younger age group of 15 to 24 years.

- Dual shoppers

Dual shoppers also like to gather product information from the Internet, but are not particularly deal prone. They like to compare brands and product features. Moreover, they tend to be single, male and in the age group of 15 to 24 years.

- E-laggards

E-laggards have a lower interest in seeking information from the Internet and tend to be in the older age group (35 years and above). However, their interest is higher than traditional shoppers.

- Traditional shoppers

Traditional shoppers do not shop online, and do not surf the Internet for comparative information, neither do look out for bargains. They buy from brick- and-mortar store and are a bit older (40-49 years). (Keng et al., 2003, pp. 149)

The respective type of online shopper influences the decision to buy or not to buy. Besides that, the online shopping determinants perceived customer value, user interface, and risk perception are important for the decision making process. They are introduced in the following chapter.

4 Approach-Avoidance Conflict in Online Shopping Situation

Applying the approach-avoidance conflict described above in the context of online shopping, approach as well as avoidance tendencies have to be identified. Real-life events are rarely influenced by only one factor. Therefore, several different dimensions with varying weight which influence the decision to buy or not to buy were identified to represent approach tendencies and avoidance tendencies. This is different to Miller’s experiments (Miller, 1944, 1964), where the only dimensions involved were hunger (induced by lack of food) and fear of pain (induced by electro shocks).

The most salient determinants found by literature review are perceived customer value, user interface as well as risk perception. Perceived customer value with its dimensions time and costs saving as well as convenience is presented in section 4.1. Following that, user interface of the website including perceived usefulness, ease of use, and effective website design is shown in section 4.2. Moreover, risk perception including its dimensions product performance risk, financial risk, psychological risk, time risk, social risk, and physical risk are essential to consider and are described in section 4.3.

For further explanation, one has to emphasise that analogously to the approach-avoidance conflict only those determinants are relevant which strengthen or weaken the already-taken intention to buy a certain product in an online shop. Otherwise, the person would not aim towards a specific goal.

4.1 Perceived Customer Value

The consumer perceives a certain value of a product when considering whether to buy it. According to Chen and Dubinsky, perceived customer value can be defined as a “consumers’ perception of the net benefits gained in exchange for the costs incurred in obtaining the desired benefits.” (Chen & Dubinsky, 2003, p. 326) Applied to the current context, benefits are gained by the shopping mode. In contrast to “brick-and-mortar”- shopping, saving costs, time, and efforts are the benefits of online shopping. (cf. Klietmann, 2001; Rohm & Swaminathan, 2004)

4.1.1 Costs saving

Most online shoppers think Internet retailing offers price advantages in comparison to traditional shopping due to reduced overhead costs for the retailer (Dennis, Harris, & Sandhu, 2002; Reibstein, 2002). Joines, Scherer and Scheufele (2003) have identified price as one of the primary motivations to shop online. Online shopping enables people to compare prices among a wider variety of merchants (Grewal, Iyer, & Levy, 2004, p. 706). This gives the possibility to search for the best deal and consequently creates stronger competition among retailers.

However, there is controversy about the role that price plays in the purchase process. On the one hand, customers show increased price sensitivity when online shopping. But on the other hand, there is the argument that price sensitivity should play a lesser role on the Internet than in traditional stores because there is so much other product information available which is not related to the price (Reibstein, 2002, p. 467). In this regard, Alba et al. (1997, p. 51) stated and Lynch and Ariely (2000, p. 101) empirically verified that the degree of price sensitivity depends on the product category. For rather standardised and easy comparable goods such as books, high price sensitivity is expected whereas for differentiated products like wines the easy cross-store comparison had no effect on price sensitivity (Lynch & Ariely, 2000, p. 101). Thus, price seems to play a lesser role when it comes to quality attributes.

4.1.2 Time Saving

More and more people are suffering from a lack of time. A solution to this problem is to reduce the time needed for things like shopping. Bellman, Lohse and Johnson (1999, p. 38) report that saving time is the main reason for online shopping and is seen superior to costs savings.

Contrary, Alreck and Settle (2002, p. 34) have shown that even though consumers evaluate the time savings aspect for online shopping quite positively, they do not consider it as causation. Furthermore, Dennis, Harris, and Sandhu (2002, p. 287) state that online shopping is not saving as much time as people would think because of time spent on searching. Thus, the internet skills of online shoppers might also have an impact on the duration of the shopping process.

4.1.3 Convenience

Regarding shopping motives, perceived convenience offered by online retailers is the most important factor to buy at home (Swaminathan, Lepkowska-White, & Rao, 1999). Location is becoming irrelevant. It is now possible to shop at home or at the office any time of day, any day of the week (Burke, 2002, p. 415; Rohm & Swaminathan, 2004, p. 750). In literature, shopping convenience often includes the time saved by a consumer (cf. Swaminathan et al., 1999). If one shops conveniently not leaving the house, consequently the person also saves the time to go to the shop. This might also explain why both time and convenience are stated as equal primary motivations to purchase in an online shop. According to Donthu and Garcia (1999, p. 52), Internet shoppers are more convenience- seekers than non-internet shoppers.

Moreover, the degree of convenience can also be affected negatively in case of slowness, unreliability (of both the Internet and the shopping process), and insufficient aftercare (Dennis et al., 2002, p. 287). Thus, the benefit regarding convenience is limited if one is bothered by a slow or unreliable connection or gets the wrong product delivered.

Besides perceived customer value with its dimensions costs saving, time saving and convenience, user interface is a further salient determinant of the purchase decision and is introduced in the following.