THE TEMPTATIONS OF STORE
WINDOW DRESSING:
How window display focus of brand stores encourages store entry
decisions through shopping motivations and brand familiarity.
Deirdre Geertrudis Theresia Maas (S 10 10 264) Master Thesis Business Administration Specialization: Innovation & Entrepreneurship
Radboud University Nijmegen Supervisor: dr. Nanne Migchels
Tell me where your freedom lies The streets are fields that never die
Deliver me from reasons why… - J.D. Morrison, 1967 -
ABSTRACT
Since the rise of online shopping, the transition to more experience-oriented shopping in city centers started (Sachdeva & Goel, 2015). The aim of this research is to help clothing brands with the design of the window display by focusing on different customer shopping situations. Store windows are the first impression a customer gets from a brand store and therefore assumed to influence store entry decisions. Whether this results in approach or avoidance behavior might be influenced by shopping motive and cognitive abilities (brand familiarity).
The research framework is organized based on the stimulus-organism-response model. Two studies have been conducted to answer the research question. First, a quasi-field experiment of 391 structured observations, resulted in the finding that the type of window display does not affect store entry when not taking any situational factors into account. The second study was a scenario-based experiment. Participants, aged between 18-35, were exhibited to two shopping motivation scenarios and asked to fill in their store entry decision after seeing a (manipulated) window display. Again, the influence of type of window display on store entry was still non-significant. Furthermore, the results showed that brand familiarity had a positive effect on the relationship between window display type and store entry. When controlling for high brand familiarity, there was no significant effect for type of window display on store entry. This changed when shoppers had a recreational shopping motive. Artistic window displays led in that case to stronger store approach behaviour, compared to product-focused window displays. For task-oriented shoppers, the type of window display did not influence store entry, even when controlling for high brand familiarity. Whereas shoppers often indicated strong store entry intentions towards the artistic-focused window displays, entry decisions for the product-artistic-focused window displays remained vague.
To conclude, for clothing brand stores the best strategic option is to implement an artistic-focused window display. This is substantiated by the fact that shopping entry is higher after seeing the artistic window display (vs. product-focused window displays). Especially, as a rememberable window display eases store entry decision making and in turn is also likely to increase brand store familiarity.
CONTENTS
1. An introduction to the research...1
Outline of the thesis ... 2
2. The theoretical view ...3
2.1 Window displays ... 3
2.2 Information processing & decision making... 5
2.3 Conceptual model ... 9
3. The studies ... 11
3.1 Study one: structured observations ... 11
3.2 Study two: scenario-based experiment ... 14
4. The results of… ... 20
4.1 Study one: structured observations ... 20
4.2 Study two: scenario-based experiment ... 21
5. To conclude with… ... 32
5.1 Discussion ... 32
5.2 Conclusion ... 33
5.3 Practical recommendations ... 35
References ... 38
Appendix 1: Observations: stimuli ... 44
Appendix 2: Survey design: stimuli ... 45
Appendix 3: Operationalization ... 47
Study 1: structured observation ... 47
1. AN INTRODUCTION TO THE RESEARCH
Customers rely on external and internal cues when shopping. Atmospheric stimuli can therefore have a huge impact on the success of a company (Bitner, 1990). This is especially relevant since the rise of online shopping resulted in a shift towards more customer experience-oriented retail (Sachdeva & Goel, 2015). Composing the right atmospheric stimuli is therefore a vital part of the marketing strategy of brands. One of the first impressions of a store is the window display. Window displays help shoppers to get a first glimpse of the stores’ merchandise (Mower, Kim & Childs, 2012). Thus, shoppers actually acquire information about the store from this tool, which in turn might influence approach behaviour. In that sense window displays can be seen as an advertising or branding tool (Sen, Block & Chandran, 2002). Many well-known department stores already made the shift to more experiential, innovative window display designs: “We’ve moved away from using traditional mannequins; instead focusing on alternative fashion forms and innovative new ways to showcase products. Our aim is to create captivating scenes that are more like art pieces in their own right; something that will make somebody want to stop and take a photo.’’ (Selfridges, 2020). The question remains if this strategy is useful in order to attract (new) shoppers to enter the store. Currently the most important questions with regard to window displays are: ‘’Which items should be displayed?’’ and ‘’How should we present these items?’’ (Lange, Rosengren & Blom, 2016). This research will focus on the latter question.
Literature about creativity in window displays is scarce (Lange et al., 2016; Roozen, 2019; Turley & Milliman, 2000). Besides, literature about in-store creativity is not usable due to the fact that the mental attitude of shoppers inside the store differs from outside the store (Turley & Milliman, 2000). Closest to the effect of creativity in window displays, is advertising literature. Which approves that shoppers prefer more creative advertisement compared to less creative advertisements (Modig & Rosengren, 2014; Lange et al., 2016). This can partly be explained by the fact that more creative advertisements receive more attention, and therefore increase customer experience. This is also in line with the research of Cornelius, Natter & Faure (2010) who showed that innovative storefronts influence store image positively, as they are valued as more modern.
Although many stores sell the same products; shoppers prefer to enter one store over another. This decision process depends on several extrinsic factors (e.g. the outside store environment) and intrinsic factors (e.g. customer characteristics) (Sen et al., 2002). How people process window display information depends on their cognitive abilities and motivation, which in turn leads to positive or negative thoughts and
due to the fact that earlier exposure and/or experience with a brand results in knowledge, which might make noticing and processing a window display easier (Kent & Allen, 1993). The goal of this research is to reveal shoppers’ preferred window display design, based on their motivation for shopping and familiarity with a brand. This is of importance as feelings of fit and positive emotions lead to impulsive buying behaviour and therefore also store entry (Chang, Eckman & Yan, 2010). As a result, the following research question will be answered:
Under which combination of situational variables, does the type of window display influence brand store entry decisions of shoppers?
Knowledge about the actual effects and use of window displays is scarce (Roozen, 2019). The available and sometimes outdated literature until now has mainly focused on the size and information being displayed (product vs. store related) (Sen et al., 2002; Mower et al., 2012). This research therefore will add to the stream of literature about external store atmospherics, which is part of retail marketing. Also, as this research is partly based on advertising literature, it will add to the empirical stream of creativity in advertising literature. At last, surprisingly, types of window displays have not yet been researched in combination with brand familiarity (Lange et al., 2016). Therefore, this research will also add to brand advertising literature.
Practically, by answering the research questions, managers and marketeers will have useful insights into how to effectively design the window display, which is often part of a retail marketing strategy (Arnold & Reynolds, 2003). Managers can adapt their strategy to the specific type of customer motivation they already facilitate in-store or would like to attract (Kang & Park-Poaps, 2010). Also, brand familiarity might put pressure on the innovation and communication strategy of the window displays.
OUTLINE OF THE THESIS
The first chapter is an introduction to this research. The highly relevant topics, relevance and goal of this research have been explained. The second chapter consists of theoretical explanations for the effects of types of window display designs, visual information and decision making. In the third chapter, the research methods will be clarified. These three chapters formed the preface of the research. After the studies were conducted, the last three chapters have been written. The fourth chapter will reveal the results of two studies. In the fifth chapter, a conclusion is drawn based on the results. Furthermore, the discussion will give a critical review on these results and the performed studies.
2. THE THEORETICAL VIEW
2.1 WINDOW DISPLAYS
Shopping experiences can be influenced by retailers through the structuring of environmental cues, also termed ‘’atmospheric stimuli’’ (Kotler, 1973). A fitting environment is especially important since shoppers are largely influenced by the visual dimension. Up to 40 percent change their mind because of something they see, learn, or do in a store (Court et al., 2009). Atmospheric stimuli can be divided into five categories (external variables, general interior variables, layout and design variables, point-of-purchase and decoration variables, human variables), of which the exterior of the store is relevant for this research (Turley & Milliman, 2000). External stimuli significantly have an influence on the behaviour of shoppers (Ward, Bitner & Barnes, 1992; Edwards & Shackley, 1992). For example, the presence of window displays (vs. none) leads to higher patronage intentions (Mower et al., 2012; Edwards & Shackley, 1992; Sen et al., 2002).
Also, window displays are the first impression a customer gets before entering a store. They are primarily used to acquire (decision) information about the store, its merchandise and services (Sen et al., 2002). This might in turn boost traffic and score interest of unfamiliar shoppers. Fashion retailers generate more store entry when a window display communicates stimuli that infers product fit (e.g., life-like mannequins) and store image, as opposed to advertisement & product-related information (Sen et al., 2002; Somoon & Sahachaisaeree, 2010). Store image can be translated through design factors and ambience (Baker et al., 1994). Literature about design factors in window displays shows that texts, spotted lighting, warm colours and a clear concept are preferred by shoppers (Abidin & Aziz, 2012; Sen et al., 2002; Pantano, 2016).
2.1.1 WINDOW DISPLAY DESIGN
Currently there is a focus on experience in retail, to engage with customers. Many retailers in fashion assume that shoppers’ like complex, and as a result creative, experiences (Jang, Baek and Choo, 2018). No clear definition about creativity exists, although it often consists of two components: originality and relevance (Sasser & Koslow, 2008). In this sense, creative window displays can be defined as ‘’original’’ in the use of existing materials in a new or artistic way and ‘’relevant’’ as it provides an organization with (in)direct value (Modig & Rosengren, 2014; Lange et al., 2016). I choose to follow the research of Oh & Petrie (2012) by making a distinction based on the level of creativity: product- and artistic-focused window displays. Although, other researches use the terms ‘’less vs. more creative’’ (Lange et al., 2016; Roozen, 2019). A product and an artistic focus can also be seen as conceptual opposites, a distinction which is proven
difference between the two window displays can be explained by Environmental Preference framework (Kaplan & Kaplan, 1989). This theory states that shoppers want to make cognitive maps by recognizing objects, make predictions and oversee consequences in an environment. The shoppers’ preference for an environment is based on two basic informational needs when approaching: they want to make sense (understand) and/or want to be involved (explore) in the environment. Visual elements that facilitate understanding are mostly associated with organized and concrete environments, which is the case with product-focused displays (Oh & Petrie, 2012). The focus lies on promoting merchandise in a straightforward, less creative style. Consequently, exploration is facilitated in complex (i.e. amount of information) environments. In artistic displays, implicit messages are exhibited, they are abstract, reveal the store’s image and as a result grab attention through high creativity (Oh & Petrie, 2012). Research by Sen et al. (2002) showed that window displays that expose store-related information (e.g. artistic) are more attractive than product-related information when the retailer’s goal is store entry. However, still each type of window display has their pro’s and con’s. Artistic displays tend to use not enough merchandise to communicate a clear expectancy message and product-focused displays may lead to non-attractiveness for shoppers (Oh & Petrie, 2012).
2.1.2. CREATIVITY IN WINDOW DISPLAYS
Lange et al., (2016) showed that more creative (vs. less creative) window displays are more successful in attracting store visits and that this effect is mediated by store window attitude, product beliefs, and perceptions of retailer effort. This is conspicuous, as greater creativity has proven problematic in an in-store environment but seems to be beneficial for store windows displays. Researchers suggest that this is due to the fact that the effect of window displays is akin to advertising (Lange, et al., 2016; Sen et al, 2002). There are two possible explanations for this: (1) window displays influence retailer-related cognitions, responses and decisions of shoppers. And (2) activities outside and inside a store serve different functions, as the interior atmospherics helps customers to guide them through the store, while external atmospherics helps to get an impression and information about the store (Cornelius et al., 2010; Mower et al., 2012; Turley & Milliman, 2000). Shoppers thus should process store windows in a way similar to their processing of advertisements, rather than the processing of in-store displays. In retail advertising, creative ways of presenting merchandise have shown to strengthen the effects of communication. Due to the incorporation of external design object(s), the advertisements were perceived as more exciting, fashionable, and sophisticated (Modig & Rosengren, 2014). An explanation for this is that customer values ‘’are grounded in a trade-off between quality and price’’ (Modig & Rosengren, 2014. P.454). Therefore, creative
advertisements can be seen as a quality cue. Shoppers perceive creative companies as more competent, which leads to positive (brand) attitudes, approach behaviour, purchase intentions and positive quality perceptions (Modig & Rosengren, 2014; Cornelius et al., 2010; Sasser & Koslow, 2008;). Besides, creative communication increases attention and the processing focus of shoppers (Pieters, Wedel & Batra, 2010). Based on literature, artistic window displays seem to have a slightly more positive effect on store entry. This leads to the following hypothesis:
H1: Artistic-focused window displays have a positive, and stronger effect (vs. product-focused) on store entry decisions, regardless of situational factors.
Although, it seems that literature has shown that more creative window displays are highly beneficial, one research pointed out that creative versus non-creative messages in window displays performed the same (Oh & Petrie, 2012). However, the same research shows that when a shoppers’ cognitive load is high, creative displays seem to be less beneficial. Since prior researches showed occasionally contradictory results, it can be stated that the actual effects of the types of window displays is likely to depend on combinations of ‘’various characteristics of the customer, the product category, the retail context, and the shopping task’’ (Sen et al., 2002. p.277). This research will focus on the shoppers’ psychological characteristics, as cognitive processes of an organism moderate store entry decisions making. Therefore, the next paragraph will discuss the likelihood of elaboration which consists in this research of the variables: shopping motivation and brand familiarity.
2.2 INFORMATION PROCESSING & DECISION MAKING
Important for store entry decisions, are the considerations shoppers make. (Information) stimuli capture attention in an environment, which will automatically influence thoughts, feelings and behaviour. This is also explained in the Stimulus-Organism-Response model (S-O-R), which states that the factor ‘organism’ (O) mediates the effect of the environmental stimuli (S) on the shoppers’ behaviour (R-response) (Mehrabian & Russell, 1974; Babin, Hardesty & Suter, 2003; Turley & Milliman, 2000).
Theories of informational processing are often used to research shopping attitudes and behaviour. Well-known is the heuristic-systematic model of information processing (HSM), which states that high message involvement results in a systematic information processing strategy, while low message involvement leads to a heuristic processing strategy (Chaiken, 1980). Ramsøy & Skov (2014) state that information can either be processed consciously or subliminal (affective) and influences shoppers’
to different more impactful behaviour with regard to decision making, memory and preferences (Baars, 2002; Ramsøy & Skov, 2014). Unfortunately for brand stores, shoppers are nowadays overwhelmed by stimuli when walking through a city. Therefore, they only process a few preferred window displays’ consciously. However, Thomsen et al., (2002) showed that there is a link between affective and conscious information processing. The effect of emotions in affective processing namely influences the possibility that a stimulus is consciously detected, and therefore behaviour (Ramsøy & Skov, 2014).
The actual likelihood of elaboration with a window display, is determined by the individuals’ level motivation and cognitive ability. Both have to be sufficient (Petty & Cacioppo, 1986). The effect of variables like emotions and credibility differ per individual and result in different levels of motivation and abilities to elaborate (i.e. understand and think). Motivation can be determined in several ways. In this research, it will be focused on shopping motivation, as this influences the interest in window display information. Also, the ability for processing a message can be dependent on several variables. Brand familiarity, which is influenced by prior experiences, is strongly linked to window display preferences due to the existing knowledge structures.
2.2.1 BRAND FAMILIARITY
Brand familiarity can be described as the amount of (in)direct experiences with a brand or store, accumulated in the mind of the shopper (Alba & Hutchinson, 1987). Research has focused on the effect of brand familiarity with regard to reactions on advertisements, information acquisition and decision making (Johnson & Russo, 1984). All these elements are also considered related to window displays. When shoppers pass by or enter a store, their brand image perceptions are influenced by the atmospheric design (e.g., type of product, quality, ambiance) of a window display (Ramsøy & Skov, 2014), which also leads to approach or avoidance behaviour (Cornelius, et al., 2010). Shoppers might see familiar brands as being available for them, which increases attention mechanisms, likeability, preference and interest (Park & Stoel, 2005). This attention has the ability to result in conscious processing of window displays, which is in turn important as it is mostly effective for approach behaviours (Ramsøy & Skov, 2014).
Frequent exposure to branding activities leads to higher brand familiarity, due to stronger knowledge structures (Kent & Allen, 1993). Strong brand familiarity in turn facilitates new information learning and makes accessing alternatives easier. Alternatives are easier processed because previous knowledge about relationships between optional brands already exists. Only relevant information will receive attention and will be analysed for decision tasks (Johnson & Russo, 1984; Kent & Allen, 1993; Ramsøy & Skov, 2014).
Moreover, a brand name is also seen as one of the most important product quality cues, that affect shoppers’ purchase decisions and will therefore likely influence store entry decisions as well (Grewal, Krishnan & Baker, 1998; Ailawadi & Keller, 2004). According to Johnson & Russo (1984) brand familiarity can be seen as an effective customer segmentation technique, which needs be used to match with the marketing-communication strategy. Window displays are part of such a strategy and provide information through atmospheric stimuli for shoppers.
Based on Herzog, Kaplan & Kaplan’s (1976) research on environmental preference theory and familiarity in urban places, it is assumed that shoppers process specific information in an environment based on preference and their familiarity with an environment. Johnson & Russo (1984) also mention that ‘’the provision of information to customers is clear: presentations should be limited to the attributes most relevant to preference judgments’’ (p. 549). I assume that due to previous research outcomes, artistic window displays are currently more preferred, compared to product-oriented window displays (as stated in hypothesis 1). Especially, as retail is changing into a more experience-oriented environment, artistic window displays will better fit into this trend (Sachdeva & Goel, 2015). The following hypothesis is developed:
H2A: Brand familiarity influences the effect of window display type on store entry decisions.
If shoppers are more familiar with a brand, cognitive mapping will become easier due to prior knowledge. Brand familiar shoppers therefore have the ability to easier process irrelevant information and make more trustful associations for decision making. As a result, artistic-focused (explorative) window displays will encourage approach behaviour decisions due to feelings of comfortability with the brand and environment (see table 1). Besides, high levels of entertainment are integrated. On the opposite, when brand familiar shoppers are confronted with a product-focused window, they might get bored of the visual display since information and expectations are already available in their mind. Arousal levels, which have proven to be important for approach behaviour, will decrease (Kaltcheva & Weitz, 2006). Store entry decisions will in that case be negatively influenced. This leads to the following hypothesis being proposed:
H2B: High brand familiarity has a more positive influence on the effect of artistic-focused window displays on store entry decisions (COMFORT), compared to product-focused window displays on store entry
decisions (BORING).
x Low Preference High Preference
2.2.2 SHOPPING MOTIVATION
The focus on shopping motivation in consumer behaviour research has increased in the past few years (Wagner, 2007). Especially as shopping motivation is a potential for customer segmentation and therefore can be used for marketing strategies (Westbrook & Black, 1985). Several researchers tried to develop taxonomies of customers by linking customer types to principles of shopping motivations. The most widely recognized human motivation literature is McGuire’s (1974) 16 types of fundamental motivations, which focusses on motives based on cognition and affection. Furthermore, Tauber (1972) distinguished three categories of shopping motivation: personal, social and impulses. Building on both McGuire’s (1974) and Tauber’s (1972) motivational taxonomies, the term ‘’shopping’’ is not basically the acquisitions of goods. Satisfaction of needs and goal attainment which are not related to product acquisition can be a motivational influence for shopping as well (Westbrook & Black, 1985). Kaltcheva & Weitz (2006) compared different labels of shopping motives, which resulted in defining two underlying shopping motives: task-oriented and recreational shopping motives. The latter can be described as the affective response, including fantasy and emotive aspects of consumption (Haas & Kenning, 2014). In that sense, shopping can be seen as generating a satisfying emotional experience instead of product consumption (Babin, Darden & Griffin, 1994). It can be assumed that shoppers with a recreational shopping motive is significantly larger now, since the emergence of online stores changed the retail landscape into more experience-oriented landscape (Gregg et al., 2017; Sachdeva & Goel, 2015). This is substantiated by the fact that lunch cafes seem to pop-up everywhere, as well as more experience oriented (flagship) stores in cities.
A shopper with a recreational motive, is likely to have a different mindset which is more open for creative window displays. Evident is that external objects are easier to process for shoppers with abstract goals. Also, the inclusion of explorational information, which is related to artistic-focused window displays, is more desirable for recreation seeking shoppers as it can be perceived as entertaining. Especially, concerning the fact that recreational shoppers, differ in their (longer) time spending and needs from task-oriented shoppers (Roozen, 2019). Task-task-oriented shoppers will more likely prefer straightforward and understandable information. To conclude, I propose that high familiar shoppers (high brand knowledge structures) with a recreational task (low information needs) have already sufficient information and therefore prefer more entertainment –oriented, artistic-focused window displays. As a consequence, I propose that task-oriented shoppers do not prefer stimulating, creative atmospherics. Shoppers with a concrete task have the desire to find specific and understandable information in window displays. Complex, multisensory atmospheres might create confusion. Recreationally motivated shoppers in turn prefer this and are therefore less affected. H2B proposed that brand familiar shoppers might experience boredom
when confronted with product-focused window displays. However, boredom might vanish as shoppers with a concrete shopping goal seem to prefer extra information about the products being available, instead of the experience of the window display. In sum, the following two hypotheses are proposed:
H3A: Product-focused window displays have a positive effect on store entry decisions, when shoppers are highly brand familiar and have a task-oriented motivation (+).
H3B: Artistic-focused window displays have a positive effect on store entry decisions, when shoppers are highly brand familiar and have a recreational motivation (+).
The before mentioned S-O-R model can be interpreted as the foundation for the proposed relation. Frequently used mediators for ‘organisms’ are Pleasure, Arousal and Dominance (PAD), of which pleasure and mainly arousal have the most powerful effect on behaviour (Donovan & Rossiter, 1982). Arousal theory assumes that the shoppers’ level of arousal is affected by the environment. Therefore, in explorative and entertainment-focused environments, easier shopping tasks are desirable. Research showed that highly stimulating environments are achieved by collative properties (e.g., artistic), as they encourage exploration and less understanding (Silvia, 2005). Therefore, arousal resulting from stimulating environments for recreational-orientated shoppers, will likely result in store entry and purchase intentions (Kaltcheva & Weitz, 2006).
2.3 CONCEPTUAL MODEL
The relation that will be researched is between types of window display design and approach or avoidance behaviour towards store entry. Hypothesis 1 focusses on the main relation, while hypotheses 2 and 3 focus of the relation mediated by situational variables (see table 2). As shown in the conceptual model (figure 1: visualization of the conceptual research model), the S-O-R model is applied. The goal is to find out if either one of the two opposite stimuli will more effective to increase positive store entry decisions (approach behaviour), under certain shopper-related circumstances. This process is mediated by ‘organisms’, which is responsible for the shoppers’ likelihood of elaboration with the window display. The two underlying processes are shopping motivation and cognitive ability (brand familiarity). As a result, the shopper responds with approach or avoidance behaviour towards the brand store.
Artistic-focused window & high brand familiarity
Product-focused window & high brand familiarity
Recreational motive + (H3B)
Task motive + (H3A)
No motive included + (H2B, versus) - (H2B, versus)
Table 2: Overview of hypotheses 2 & 3
3. THE STUDIES
3.1 STUDY ONE: STRUCTURED OBSERVATIONS
The goal of this research was to discover which type of window displays shoppers prefer to encourage store entry decisions, under certain cognitive conditions (motivation and brand familiarity). The hypotheses retrieved from chapter 2: theoretical framework, have been tested and led to final conclusions. Two separate quantitative studies have been conducted. The first study tested main relationship: the effect of window display type on store entry. The second quantitative study took the influence of the situational variables on the main relationship into account.
Previous quantitative researches have showed that creative or artistic-focused window displays led to positive store entry decisions (Lange et al., 2016; Roozen, 2019; Sen et al., 2002). To get a glimpse about how customers respond to window displays in situ, this study adds value to the existing literature by conducting structured observations.
In reality, not all stores in a city center use window displays. It is significantly more often seen in fashion and lifestyle stores, which are often high product-involvement goods (Sen et al., 2002). These types of stores add to the unique ambiance in city centers and are therefore important for cities (Padilla & Eastlick, 2009). Female fashion stores seemed to fit best for this research, as both types of motivation are commonly used when shopping for clothing (study 2). The observed females were guessed to be aged between 15 and 55, as this was likely the main target groups of the chosen branded stores. The structured observations took place in Nijmegen as this city centre has quite similar sized windows, which is proven to effect store entry as well (Mower et al., 2012). Also, all window displays front towards the (high) street, which eases comparability. Otherwise, external validity of the study could decrease when doing structured observations. The observations were conducted on weekdays only (Thursday & Friday) and at different times. Crowdedness was therefore a less influential factor and still enough observations were generated (Roozen, 2019). Also, as window displays often change on a weekly basis, all observations had to be done in only one-week time. It is important to mention that due to the regulations with regard to COVID-19, it is highly possible that less or different shoppers have visited the city centre. Initially, 400 observations were targeted. Although structured observations are quite subjective, it gives a perfect sample of behaviour in reality. Hence, to increase inter-reliability two independent observers noted actual shopping behaviour, so subjectivity could be reduced. Furthermore, it is argued that observations are less preferred, due to ethical
considerations. It was decided to still execute this experimental field research. Although shoppers were not informed, it would not harm anyone, and anonymity was guaranteed.
3.1.1 PROCESS OUTLINE AND OPERATIONALISATION OF THE OBSERVATIONS
Pretest of stimuli
First, it was important to conduct a pretest, to select the right female clothing window displays for observation. It was assumed that brand familiarity, price image, similarity in (daily apparel) collection, as well as store buying behaviour were quite similar for the selected displays.
Secondly, four pre-selected window displays were chosen (see appendix 1). The window displays with an artistic-focus, were chosen based on the inclusion of (the amount of) atypical objects in an unusual composition, originality and relevance (Pieters et al., 2010; Lange et al., 2016). The pre-selected window displays with a product-focus, only showed mannequins wearing the products. No or limited atypical objects were displayed.
Thirdly, a photo of the window displays was taken, so participants would not have to walk to the different window displays and their opinion could be asked quickly. In total 25 females, with a respective age of 20-40 years old, at the streets in Nijmegen were kindly asked to give their opinion about the window displays. The operationalisation of the items with regard to finding the right window displays (information) followed earlier research and was therefore based on perceived creativity level (Roozen, 2019; Lange et al., 2016). Furthermore, Oh & Petrie (2012) have shown that the environmental preference theory is applicable on the two opposite window display types (e.g. dividing between explorative = mystery & complexity, and understandable = coherence & legibility). For a complete overview of the operationalization used for this study, see appendix 3. The design of the questions for finding the right window displays was guided by the following indicators:
• Artistic-focused display: explorational information, more creative (complex). • Product-focused display: understandable information, less creative (simple).
This resulted in the following question being asked: ‘’When walking past this window display, what are your perceptions about the level of creativity in the design?’’. The scale was adapted from Roozen (2019) and ranged from 1 to 7 (1 = simple, 7 = complex). Next, the actual two window display variables used in this research were questioned by showing the window display after each other. The goal of this item was to confirm that the right selection has been made. The questions were: ''Do you perceive this window display as product-focused?'' (showing a product-focused display) or ''Do you perceive this window display as
artistic-focused?'' (in the case of artistic-focused window displays). The possible answer options were binary: yes or no. All window displays were questioned.
One day before the actual field research, the within subject design pretest was analysed. The statistical results from a paired samples t-test showed that the perceived creativity of participants confirmed that artistic window displays (WDA) were indeed seen as more complex, compared to simple perceptions (WDP) (See table 3, p =.000 for all pairs). To be more precisely, the second questions indicated that the that the variable opposites are indeed clearly represented in terms of their concept (yes = 92%). To conclude, it can be stated that the selected window displays are a good representation of the window display variables and can therefore be used for the structured observations.
Pairs Mean value (SD) t-value, (df = 24) Significance
WDA 1 x WDP 2 4.76 (1.16) & 2.84 (.80) 6.410 P = .000 WDA 2 X WDP 1 5.76 (.88) & 3.04 (.73) 10.947 P = .000 WDA 2 X WDP 2 5.76 (.88) & 2.84 (.80) 13.094 P = .000 WDA 1 X WDP 1 4.76 (1.16) & 3.16 (.73) 6.577 P = .000
Table 3: Comparison of the concept presentation artistic-focused window display (WDA) and product-focused window display (WDP): outcomes of the paired t-tests for the pretest of stimuli.
3.1.2 THE OBSERVATIONS
For the structured observation, two people observed shopping behaviour independently. Store entry and looking into the window display were noted. In total 400 independent observations took place, of which 100 for each window display (2x artistic and 2x product-focused). The observers stood at a small distance of the window display. It is highly possible that shoppers have seen them making notes, but this was impossible to overcome due to the architecture of the buildings. Due to non-crowdedness in front of the window displays, the observations were almost similarly noted. The 9 cases which were different, have been taken out for the analyses. The results of this study will be discussed in chapter 4: Results.
3.2 STUDY TWO: SCENARIO-BASED EXPERIMENT
In a second study, the hypotheses that were based on cognitions that underly the will for elaboration with window displays and therefore store entry decisions have been researched. Quantitative methods seemed to fit best as behavioural choices guided by stimuli, had to be measured.
It was important that this research tried to replicate a real-life situation as close as possible. The best option would have been to collaborate with a fashion brand and change the window display designs. However, this was impossible due to the short time frame of this research and the effort a retailer would have to put into it. Moreover, since different situations (i.e. shopping motivations) had to be compared, there would not be enough information about the situational variables to answer hypotheses 2 and 3, when only observing actual behaviour (study 1). Therefore, the hypotheses have been tested by executing a scenario-based experiment. Many critics believe that scenario-based experiments, due to the shortage of a real-life context, produce unrealistic data of behaviour (Bardsley, 2005). However, a scenario-based research does ease the use of manipulating and controlling variables, which leads to an increase of the internal validity of the research (Kim & Jang, 2014). Furthermore, scenario-based experiments are especially useful when measuring concepts based on cognitive evaluations, as is the case in this research (Kim & Jang, 2014). Two different scenarios (manipulations) have been constructed with a 2 (window display type: artistic/product-focus) x 2 (motivation: task-oriented/recreational) x 1 (brand familiarity) design.
Shopping motivation in the scenario-based experiment was asked from a personal decision-making perspective. Roozen (2019) argued that including questions from a third-person perspective for the task-oriented scenario increases the validity of the research. The reason behind this is that participants with a specific task might not decide to enter a store, due to personal brand opinions (e.g., preference) (Fisher, 1993). Since this research included brand familiarity, I had chosen to ask both motivational scenarios from a self-perspective. Furthermore, some research proposes that there is a gender bias on shopping orientation and buying behaviour (Gam, 2011). This research focussed on female shoppers only, so validity could be assured. Lastly, I expected all participants to have prior experience with both types of shopping motivation before and as a result were able to emphasize with the scenarios.
3.2.1 PROCESS OUTLINE AND OPERATIONALISATION OF THE EXPERIMENT
Pretest of stimuli
A pretest was carried out to select window display stimuli that met the theoretical explanations. The method was similar to the selection of window displays in the first study: based on the participants’ perceived
creativity level and concept validation of the window display (Roozen, 2019; Lange, 2016). As brand familiarity was now included as a covariate or control variable, sufficient participants with moderate - high brand familiarity levels needed to be guaranteed due to the proposed hypotheses. Therefore, the main street fashion retailer and brand: Topshop was carefully chosen. Another assumption for this brand choice was the availability of pictures of two opposite types of window displays to use as for the manipulation of stimuli.
During the preselection of window display pictures, attention was paid to conceptual opposite designs and the inclusion of the brand name. The merchandise being shown in both displays was slightly different as this could not be controlled. However, the used colours were quite equal (highly colourful). The pictures did not include clear price tags or promotion stickers, to assure that the focus was on the design being displayed.
In total 25 females were questioned for their opinion. They were all closely connected to the researcher and aged between 20-30 years old. A paired samples t-test confirmed that complexity is better represented in the artistic-focused displays, compared to simple perceptions (t(25)= 12.25, p =.000; Martistic = 6.20, Mproduct = 3.16). The other questions with regard to concept validation, indicated the window display pictures clearly represented the concepts (yes = 86%). Therefore, the selected pictures have been used for the scenario-based experiment.
See appendix 2: Survey design, for more extensive brand information about Topshop, as well as the used pictures of the window display stimuli.
The experiment
The online experiment started with an introduction about the scenario-based experiment and an explanation of ethical codes, so participants still had the option to quit participation. Furthermore, participants were informed about a price they could win when participating. After this the actual experiment could start.
First, participants had to take a good look at the independent variable: a randomized window display picture (either artistic or product-focused). Next, the first moderating variable: a task-oriented shopping motive was explained. Task-oriented shopping means that products need to be purchased. In this case: clothing. For this scenario, shopping itself does not bring satisfaction to the participant. It is purely focused on the fulfillment of the shoppers need and economic transaction. The opposite holds for the second treatment: recreational shopping. This scenario describes that a shopper freely chooses to engage
First, the constructed task-oriented scenario, based on Kaltcheva & Weitz (2006) was shown: ‘’Tonight, is your friend’s birthday party. You walk past the shown window display and all of a sudden you pay attention to the big hole in your jumper. NOOO, not today! It’s probably because that crazy cat of yours scratched you. You decide to buy a new, daily-wearable and fashionable jumper for tonight’’.
After reading this, participants were asked for the likelihood of store brand entry based on the exhibited window display and with the task-oriented motive kept in mind. Three items on a seven-point Likert-type scale (1 = strongly disagree, 7 = strongly agree) have been used to measure store entry decisions based on the window display (Lange et al., 2016). The following three items were questioned: “This store window display makes me want to enter the store to find a jumper’’, ”The products in this display make me want to enter the store, to find a jumper” and “I would probably enter the store, to find a jumper” (Lange et al., 2016. p. 1018).
Second, the constructed recreational scenario, based on Kaltcheva & Weitz (2006) was shown: ‘’It is Saturday, and you are bored at home. A friend calls you if you want to join her to the city, just to wander around a bit. You decide to join and visit some stores to relieve the sense of boredom’’.
Again, participants were asked to indicate their store entry intentions based on the three before mentioned items: “This store window display makes me want to enter the store,” The products in this display make me want to enter the store,” and “I would probably enter the store” (Lange et al., 2016. p. 1018).
Third, the level of brand familiarity (control & covariable) was measured. First, a semi-open question with regard to brand recognition was asked to get a small glimpse of confirmation with regard to high brand familiarity and to check whether the participants paid attention to the window display picture: ‘’Did you see which brand was being displayed in these window displays?''. The answer possibilities were no & yes, I saw the brand…(open).
Brand familiarity was measured by three items on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). This research followed the scale of Kent & Allen (1994) for measuring the role of brand familiarity in the effects of consumer memory in advertising. The following items were asked: (1) I am experienced with Topshop, (2) I have knowledge about Topshop, and (3) I am familiar with Topshop. The items gave insights in the level of familiarity with a brand and how this influences their cognitive process when making a store entry decision. Furthermore, brand perceived quality is perceived to be an indicator of brand familiarity (Dursun et al., 2011). It has proven to result in products being purchased more often, therefore it was included in the questionnaire as well (Ailawadi & Keller, 2004). The goal is of this research was not to explicitly research purchase intention, yet the participants’ answers gave logically an indication of approach or avoidance behaviour towards a store as well. Three statements with regard to brand quality have been
used for this research: (1) The products of store brand Topshop have a high quality, (2) The products available in the Topshop store are reliable/trustworthy, (3) The products available in the Topshop store will give me the result I'm looking for (Garvin, 1984; Dodds, Monroe & Grewal, 1991).
Fourth, the participants’ self-perspective about their buying behaviour was questioned (e.g., amount of money spent on clothing, frequency of shopping, actual shopping motivation). Shopping frequency was an important question as: shoppers who more often visit a particular brand store during their retail safari might interact less and become less sensitive for window displays (Lange et al., 2016). This is in contrast to less regular shoppers who occasionally visit the brand store, and use window displays as a guidance stimulus. In the proposed logic, this might moderate the effects of different types of window display on store entry (Sen et al., 2002). The effect is also likely to depend on the participants’ own shopping motivation. Usually, recreational shopping leads to longer shopping duration (Roozen, 2019). Therefore, participants were asked to indicate if they are more often task-, neutral- or recreational shoppers, on a 7-point opposites scale (1= task-oriented motive, 7= recreational motive). Furthermore, it seemed important to ask purchase intention directly as well: ''I would consider buying clothing from Topshop'', as biased results due to store entry avoidance could be reduced.
Lastly, general questions about age, gender, education and demographics have been asked. The living area of the participant comprised important information, as participants who life in an urban area might have more or different experience with shopping compared to village areas. All items were rated on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree).
See appendix 3: Operationalization, for an overview of the operationalized variables used for this study.
3.2.2 PRETEST OF THE EXPERIMENT
To make sure that the experiment was properly composed, a pre-test was carried out. Ten relatives and friends of the researcher were asked to fill in and criticize the questionnaire. The asked female respondents were aged between 18-55 and all had a high educational level. Due to the different ages and different work fields of the participants, understanding of the questions was assured. The feedback received, led to small alterations in the formulation of questions and the explained oriented scenario. Primary, the task-oriented scenario was written with a student focus, forgetting that most 55-year-old females do not live in a student house anymore.
3.2.3 PARTICIPANTS, SAMPLING AND ETHICS
When executing a scenario-based experiment, homogeneous groups of participants should be generated to execute a statistical analysis. According to Hair, Black & Babin (2014) the minimum size of each group in the experiment should be 30 participants, before any analysis can be conducted. Larger groups will help making results more specific, but on the other hand increases the change of false rejection due to small characteristics. As two manipulations have been carried out, at least 60 participants should have given valid answers. According to Field (2016), as a rule of thumb, each predictor should have 10-15 cases. In this case 50-75 valid answers would have been enough. To make sure that all answers were valid, it was best to generate some extra participants so invalid answers could be deleted. Therefore, the goal was to collect 100 participants. Finally, this resulted in 105 completed questionnaires. One invalid questionnaire was removed, which indicated a clear patron (same answer). The time spent and IP-addresses looked fine. Furthermore, after seeing the results of the item ‘age’, it was decided to take 6 questionnaires out of the data and change the age of the target group to 18-35 years. This was before the experiment determined as 15 – 50-year-old females, as this is the target group of Topshop according to designer I. Markham (van Rhee, 2015). The change in age seemed important, as it made the results of the research more specific. A possible reason for the age shortages is probably the network of the researcher. This alteration did not seem to be a problem, as the target group mentioned by Topshop is quite broad. I expect that the currently used target group for this research is a more adequate example of reality. Of the removed 6 answers, only one was able to reproduce the stores’ name and brand familiar. This open question about the brand being displayed, was asked after seeing the window display and answering store entry questions. 10,2% of the participants, were unable to mention the brand name. This might mean that people did not focus enough on the window display picture or the picture not being clear enough, but these answers might still be valid. However, a first glimpse about (apparently high) brand familiarity was also indicated with this question. When analysing the brand familiarity items of these participants, it was obvious that 10 of them were also unfamiliar with the brand (average item mean score of 1 or 2, on 7-point Likert scale). To conclude, in total 98 valid answers have been used for this experiment. Although a little less than the intended 100 questionnaires, it was still enough to draw conclusions from.
The experiment was online from the 19th of October till 2nd of November. Different methods of data
collection have been used to collect participants: through spreading the questionnaire on social media websites (e.g., Facebook, LinkedIn and Instagram), Surveyswap and direct messages in mobile applications. It was quite hard to find enough participants, as (in the first stance) only females between the age of 18-55
were targeted. After 10 days only 20 questionnaires were missing (out of 100). Therefore, I questioned shopping females at city the center to fill-in the questionnaire. This was in a few cases done through immediate questioning and showing the accompanied pictures of window displays on a tablet. I have asked my network to spread the questionnaire as well.
The online quantitative survey application Qualtrics was used to fill in the questionnaire. This online tool made it easier to gather a large number of responses in a tight time frame. According to Forza (2002) the will to participate in questionnaires is decreasing. Therefore, participants have been lured to participate in the research by raffling Fashioncards worth 15,- euros for those who are interested. Furthermore, the name of the Radboud University and the results were offered to assure legitimacy to participants.
The most influential limitation for the performance of this study was the fact that in an online scenario-based experiment, it is harder to control external elements. External validity of the research and unsystematic variation could have been decreased, as normally participants in a similar context are preferred. Although the best option to would have been to ask questions in a fully controlled environment, this was not possible due to the timeframe and budget of this research. Furthermore, collecting a large amount of people in a room was impossible due to COVID-19. As an attempt to solve this issue, the results of the control question for living area were analysed. The results led to the conclusion, that primarily living area did not influence the results. A valid reason might be that the experiment contained scenario-based questions in which the window display environment was similar for every participant.
With regard to ethical considerations in this research, the five APA principles for research ethics were followed (D. Smith, 2003). Participants were first questioned if they voluntarily wanted to participate in this research and informed about the possibility to end the questionnaire at any moment, if they did not feel comfortable. The confidentiality of the results and participants’ anonymity were also mentioned. Secondly, participants were informed about the purpose and expectations of this research, which is considered not to be an unethical subject. To assure the questionnaire would not be entangling for the participants, it was also important that the questions were not psychologically provocative. Therefore, the questions were asked in a short, simplistic and straightforward manner. Though a slight sense of humour was used to motivate participants to finish the questionnaire. Furthermore, the questions were asked in Dutch to make the experiment better accessible, increase validity and make sure only Dutch people participated. Lastly, the expected duration, (possibility to win a) reward and contact details were presented.
4. THE RESULTS OF…
4.1 STUDY ONE: STRUCTURED OBSERVATIONS
To find support for hypothesis 1: ''Artistic-focused window displays have a positive, and stronger effect (vs. product-focused) on store entry decisions, regardless of situational factors.'', structured field-observations are conducted. As mentioned before, the observations were almost similarly noted by the two observers. Only females, who were guessed to be aged between 18-55 years old were noted. To analyse the results of the structured observations, an independent samples t-test was applied. With this test, the effect of the independent variables: window display types, could be compared and their single effect can also be interpreted separately. First, the dependent variable ‘obvious look’ for both window display types was calculated. The Levene’s test for equality of variances was non-significant (p=.071). The results of the t-test showed that there was no significant difference between both types of window displays that draw the attention of people passing by (t(391)= -.905, p=.366). The same test was conducted with store entry as the dependent variable. The Levene’s test for homogeneity of variances was significant (p=.015), therefore equality of variances was not assumed. The results of the t-test showed that also for store entry, there is no significant difference between the two types of window displays (t(391)= -1.216), p=.225). As can be seen in table 4, still stores with product-focused window displays had less shoppers looking obviously into the window display (22,8%), compared to artistic-focused (26,8%). The same holds for store entry (8,8% vs. 5,6%), although both not being statistically different. Therefore, hypothesis 1 is only partially rejected.
Artistic-focused window display Product-focused window display Significance (p=<.05) Walking by 194 = 100% 197 = 100% N/A
Obvious look into the window display
52 (26,8%) 45 (22,84%) p= .366
Store entry 17 (8,76%) 11 (5,58%) p= .225
4.2 STUDY TWO: SCENARIO-BASED EXPERIMENT
4.2.1 DESCRIPTIVES
In this chapter the results of the scenario-based experiment are presented. A full overview of the descriptive results is shown in table 5. Two versions of the questionnaire have been distributed. The difference between the versions was the store window display: 51 participants indicated their store entry decisions for the product-focused window display, 47 participants indicated store entry for the artistic-focused window display. The average age of participants was 26.55 years old. Most participants had a scientific university level (53,1%), which might have biased the results. Advertisement literature suggests that the shoppers’ cognitive ability, leads to different responses to advertisements (G.E. Smith, 1996). For lower educated people, it might be more difficult to remember advertisements (brand familiarity) and they more often process advertisement information heuristically (Bettman & Park, 1980). Also, lower educated people are easier influenced by environmental cues, which might have led to different results. Therefore, this research should be interpreted for their effect on higher educated people only.
Participants lived mostly in urban areas (86,7%). As mentioned before, this was an important question as people who live in village might have less experience with both types of window displays of branded stores and therefore increase unsystematic variation of the experiment. It has been considered to take the 13 questionnaires of participants who lived in a village, out of the data. Taking a closer look at the data, the individual scores seemed normally distributed. Furthermore, although literature might say so, I expect that people living in a village area frequently visit city centers as well. Therefore, it is likely that this had no influence the results.
Participants were also questioned to give an indication about their buying behaviour. Most participants went at least once a month to a city center to shop for new clothing (65,3%). From this it can be concluded that offline shopping is still relevant. However, it is not questioned whether the money they spent per quarter, was on online or offline purchases. Also, participants who were confronted with the product-focused display (vs. artistic-product-focused) shopped less often. This might have had an effect on the results. As mentioned before, less regular shoppers might namely more rely on window displays as a guidance stimulus (Lange et al., 2016). As product-focused window displays are more focused on presenting merchandise in a straightforward way, it therefore fitted with the information being presented. Results should therefore not be heavily affected. Especially when combining this with the participants self-perception about shopping
motive. Recreational shoppers shop at least once a month (78,13%). While shoppers who perceive
themselves more as task-oriented, often shop quarterly (48,15%).
Questions with regard to the brand Topshop were included to control for their effect on the study. Both purchase intention and brand perceived quality might have influenced the decision to approach or avoid a brand store. First, an indication towards brand purchase intention was gathered through direct questioning. On average participants indicated that they considered buying clothes at Topshop (M = 4.93, SD = 1.66). In first stance it can be concluded that, a big influence of negative attitude towards the brand and store entry is unlikely.
Furthermore, participants perceived Topshop as medium to high quality (M = 4.71, SD 1.10). When creating a split-file for perceived brand quality level, high (vs. low) perceived quality had a significantly stronger effect on perceived purchase intention (t(96)= 25.604, p=.000; Mhigh = 5.29, SD = 1.45 & Mlow = 3.33, SD = 1.61). This is in accordance with the research of Ailawadi & Keller (2004), who also found a positive effect of brand perceived quality on purchase intention. Therefore, the inclusion of the brand name did not negatively influence the results. As expected, participants significantly considered buying clothing at Topshop more often when their brand perceived quality was high (t(96)= -.5060, p=.000; MBPQhigh = 5.29, SD = 1.45 & MBPQlow = 3.33, SD 1.61) and also when they were highly familiar with Topshop (t(96)= -5.678, p=.000; MBFhigh = 5.53, SD = 1.44 & MBFlow = 3.79, SD = 1.45). Previous brand experience is therefore an important factor during shopping. As all seem to have medium-high scores, this research should only be interpreted for high brand familiar shoppers only. As this is also stated in the hypothesis, this should have had no influence on the results.
Table 5: Overview of descriptives, in numbers and percentages. Product-focus (51=100) Artistic-focus (47=100) Total (98=100) Age 18-35 M (26.1), SD 3.06 M (27.04), SD 3.00 M (26.5), SD 3.05 Educational level Lower education Mbo Hbo Wo 0 3 (5,9%) 16 (31,4%) 31 (62,7%) 0 5 (10,6%) 22 (46,8%) 20 (42,6%) 0 8 (8,1%) 38 (38,8%) 52 (53,1%) Living area Urban area Village area 40 (78,4%) 11 (21,6%) 45 (95,7%) 2 (4,3%) 85 (86,7%) 13 (13,3%) Quarterly spent > clothes
€ < 125 € 125-250 € 250-500 € >500 11 (21,6%) 29 (56,9%) 9 (17,6%) 2 (3,9%) 6 (12,8%) 29 (61,7%) 10 (21,3%) 2 (4,3%) 17 (17,3%) 58 (59,2%) 19 (19,4%) 4 (4,1%) Shopping Frequency Once a week Once a month Several times month Several times quarter Several times year
0 15 (29,4%) 16 (31,4%) 13 (25,5%) 7 (13,7%) 5 (10,6%) 11 (23,4%) 17 (36,2%) 13 (27,7%) 1 (2,1%) 5 (5,1%) 26 (26,5%) 33 (33,7%) 26 (26,5%) 8 (8,2%) Shopping Motive (1= task-oriented, 7= recreational) M (4.71), SD 1.62 M (4.70), SD 1.91 M (4.70), SD 1.75 Purchase intention Topshop M (4.78), SD 1.65 M (5.09), SD 1.67 M (4.93) SD 1.66
4.2.2 FACTOR ANALYSIS
As this experiment questioned multiple-items, dimensions need to be formed through factor analysis. First, six items with regard to store entry were added into one factor analysis. Two factors have been identified with 85,21% cumulative explained variance. The Kaiser-Meyer-Olkin (KMO) test for sample adequacy had a value of 6.84 and Bartlett's Test of Sphericity for sufficient correlation between the sets of scale items, was significant (p=<.000). Furthermore, the communalities (>.20) and the factor loadings (>.30) were sufficient. No items were deleted. In table 6 the factor loadings, means and communalities are presented.
Item Mean SD Factor loading
Component: Communalities 1 2 ENTRYTM1 4.19 1,62 .922 .855 ENTRYTM2 3.82 1,70 .805 .706 ENTRYTM3 4.53 1,72 .872 .770 ENTRYRM1 4.84 1,69 .928 .868 ENTRYRM2 4.20 1,64 .839 .728 ENTRYRM3 5.14 1,53 .828 .711 Composed variables ENTRYTM 4.18 1.55 ENTRYRM 4.73 1.49 ENTRYINTOTAL 4.45 1.21
Table 6: Factor analysis for store entry with a task (ENTRYTM) and recreational (ENTRYRM) motive. Extraction method: Common Factor Analysis, Varimax with Kaiser Normalization, loadings below 0.40 are not shown. Total variance explained = 84,70%; KMO = .676; Bartlett’s test chi-sq. = 417.415, df = 15, p =<.000. Scale min. 1, max. 7. 4= neither agree, nor disagree & 5 = more or less agree.
A single factor analysis was performed for the items with regard to brand familiarity (BF) and another one for perceived brand quality (BPQ). The items with regard to brand familiarity had a KMO of .724 and Bartlett's Test of Sphericity was significant. The items relating to brand perceived quality had a KMO of .721 and Bartlett's Test of Sphericity was also significant. In both factor analyses, the communalities (>.20) and factor loadings (>.30) were perfect. Again, no items were deleted. In table 7 and 8 the factor loadings,
communalities and explained variance are presented.
When conducting a factor analysis with all branding-related items, two factors were indeed confirmed. However, the item ‘I have knowledge about Topshop’’ loaded on both factors. A reason for this might be the strong conceptual overlap. For example, item BPQ1: ''The products of store brand Topshop are reliable/trustworthy'', can be seen as an indication of brand knowledge which was questioned in item BF2: ''I have knowledge about Topshop'' (Dursun et al., 2011). The option to take the double loading item out of the analysis was rejected, as brand perceived quality was not used in any of the hypotheses. It is decided to not use this factor for further (extra) analyses.
Item Mean SD Factor loading Communalities
BF1 5.15 1.60 .798 .637
BF2 4.22 2.04 .948 .898
BF3 3.75 1.90 .840 .706
Composed BF 4.38 1.68
Table 7: Factor analysis for brand familiarity (BF). Extraction method: Common Factor Analysis, loadings below 0.40 are not shown. Total variance explained = 82,79%; KMO = .724; Bartlett’s test chi-sq. = 179.659 df = 3, p =<.000. 7-point Likert scale: min. 1, max. 7. 4 = neither disagree, nor agree.
Item Mean SD Factor loading Communalities
BPQ1 4.83 1.03 .875 .765
BPQ2 4.62 1.60 .736 .541
BPQ3 4.68 1.10 .863 .744
Composed BPQ 4.71 1.10
Table 8: Factor analysis for: Brand Perceived Quality (BPQ). Extraction method: Common Factor Analysis, Loadings below 0.40 are not shown. Total variance explained = 78,56%; KMO = .721; Bartlett’s test chi-sq. = 140.007, df = 3, p =<.000. Scale min. 1, max. 7, 5 = more or less agree.
All scales had a Cronbach’s Alpha between .834 and .907, therefore they were identified as reliable (>.70) (Hair et al., 2014). Furthermore, in none of the cases the Cronbach’s Alpha could be improved, as they were already quite high. Table 9 presents an overview of the Cronbach's Alpha for each variable.
Scale Number of items Cronbachs's Alpha
Entry Task-oriented Motive 3 .907
Entry Recreational Motive 3 .906
Brand Familiarity 3 .893
Brand Perceived Quality 3 .834
Table 9: Calculated Cronbach’s Alpha per variable.
4.2.3 TEST OF HYPOTHESES
In this section, the stated hypotheses will be rejected or accepted, based on the statistical test results. Furthermore, only t-tests and ANCOVA techniques are applied, as there only two types of window displays being manipulated (categorical) and store entry was measured on a 7-point Likert scale. Several researchers propose that Likert scale variables, which are usually treated as ordinal variables, can also be used as continuous variables when the scale contains 5+ options without bringing violence to the analysis (Norman, 2010; Johnson & Creech, 1983). Therefore, the ANCOVA assumption of the dependent variable to be of an interval or ratio level is met. To analyse the data, it was important that the collected data was checked on the homogeneity of the groups, sample sizes, missing items and distribution of the items. These checks are important as they give an indication if the collected data can be used for the statistic method (Hair et al., 2014).
Main effect
The results of the structured observations (study 1) already rejected hypothesis 1: ''Artistic-focused window displays have a positive, and stronger effect (vs. product-focused) on store entry decisions, regardless of situational factors.''. Due to curiosity about the results, the data from the scenario-based experiment was used to test hypothesis one again. Similar to study 1, an independent t-test was applied, with the types of window display (independent variable) and store entry (dependent variable). The Levene's test for equality of variances was non-significant (p=.869). The result of the t-test in the second study confirmed the results of study 1, there is no significant difference between of both types of window displays (t(96) = -1.915, p=.059). Furthermore, the average store entry for a product-focused display (Mproduct = 4.23, SD = 1.20) was indeed lower compared to the artistic-focused display (Martistic = 4.70, SD = 1.19). With a mean value score closest to 5 (agree): Artistic-focused window displays lead to store entry decisions, regardless of situational factors. For the product-focused window display, the average mean score was closer to 4
