Master thesis
Happy shoppers or frustrated shoppers? The effect of
store tidiness on shopper frustration.
Lisette Verheij – s1911503
2014/2015
Master thesis:
Happy Shopper or Frustrated Shopper?
The effect of in-store tidiness on shopper frustrations.
Name: Lisette Verheij Student Number: 1911503 Address: Javastraat 87a
2585 AG Den Haag E-mailadress: lmverheij@gmail.com l.m.verheij@student.rug.nl Telehone number: +31 (6) 43 67 87 73 Date: 12-01-2015 Department: Marketing
Study track: Marketing Management & Marketing Intelligence 1st Supervisor: dr. L. Lobschat
Management Summary
This research investigates the effect of in-store tidiness on customer frustration, and whether this relation is influenced by the store’s price image and sales.
Tidiness is a relevant aspect that each shopper experiences during its shopping trip, in the supermarket, and/or when buying clothing. However, not all stores are capable of keeping the store tidy. Within this research, the effect of the tidiness of shops will be investigated. It is expected that the shopper does not prefer an untidy store, and even get frustrated. These negative emotions that appear during a shopping trip could lead to avoidance behavior and should therefore be better understood and eliminated. In addition, the effect of price image and sales will be researched. It could be the case that the customer is more frustrated when a more expensive store is untidy. In advance of that, a look will be taken whether price promotions or sale have an influence on this overall perception.
The data is collected by means of an experiment. The hypotheses are tested via a 2x2x2 factorial design, so in the survey 8 different situations are shown. First, a factor analysis has been performed. Then, an ANOVA has been done to test the groups. A multiple regression analysis is done to test the relations between the variables and to test directions hypothesized.
The results reveal a strong effect between an untidy store and frustration. Furthermore, the price image plays a significant role. The higher the price image of the store, the more the consumers become frustrated in an untidy store. When a store has a sale, the relation becomes weaker; so the customer becomes less frustrated during a sale when a store is untidy and expensive.
Table of Contents
Master thesis: ... 2
Management Summary ... iii
1. Introduction ... 3
2. Literature ... 8
2.1. Emotions ... 8
2.2. Emotions and Environmental cues ... 10
2.3. Price strategies ... 13
2.4. Discounts and Promotions ... 14
2.5. Control variables ... 16 2.5.1. Mood... 16 2.5.2. Sales Proneness ... 16 2.6. Conceptual Model... 17 3. Data ... 18 3.1 Data collection ... 18 3.1.2. Obtaining respondents ... 229 3.2. Recoding ... 19 3.3. Descriptive ... 20 3.3.1. Gender ... 21 3.3.2. Age ... 21 3.3.3. Education ... 22 3.3.4. Residence ... 23 4. Methodology... 25 4.1. Factor analysis ... 25 4.1.1. Validity ... 25
4.1.2. Factor extraction method ... 26
4.1.3. Determining the factors ... 26
4.1.4. Assigning the factors... 26
4.3 Multiple regression ... 26
4.3.1. Validation for ANOVA and Regression ... 27
4.3.1.1. R-square ... 27
4.3.1.2. Levene’s Test of Equality of Error Variances (ANOVA) ... 27
4.3.1.3. Multicollinearity (Regression) ... 27 4.3.1.4. Residual diagnostics ... 28 4.3.1.5. Non-Normality ... 28 4.3.1.6. Heteroscedasticity (Regression) ... 28 5. Results ... 30 5.1. Control variables ... 30 5.2. ANOVA ... 31 5.2.1. Validation... 32 5.3. Regression ... 33 5.3.1. Model selection ... 35 5.3.1.1 R-square ... 35 5.3.1.2. Information Criteria ... 35 5.3.2. Validation... 36 5.3.2.1. R-square ... 36 5.3.2.2. Multicollinearity ... 36 5.3.2.3. Residual Statistics ... 37 5.3.2.4. Non-normality ... 38 5.3.2.5. Heteroscedasticity ... 38
5.3. Interpretation of the results ... 39
6. Discussion & Conclusions ... 41
6.1 Discussion ... 41
6.2. Conclusions ... 42
6.3. Limitations & direction for further research... 42
7. Managerial Implications... 44
References ... 46
Appendix A – Questionnaire ... 52
1. Introduction
The surge of online web shops for clothing has given the customer a new option; he or she can now choose to either buy clothing online through the internet or in the traditional offline stores. Currently it is possible to see a trend in which the sales of online shops are growing more rapidly than the sales of traditional (offline) stores. The majority of sales are still made offline, nonetheless, the market share of online shops is increasing compared to offline stores (Unruh-Enos, 2014). This growth in online shopping can be explained by, for example, convenience, the ability to search for information and products, and price. (Ganesh et al, 2010). To not lose further terrain to the online stores, traditional shops are reinventing themselves to offer new and unique shopper experiences for their customers. By changing the design of the store or ambient cues, stores are trying to differentiate and attract more customers.
Abercrombie & Fitch is a well-known example for creating an outstanding, unique shopper experience. Abercrombie is perceived as a rather expensive clothing shop. This price image is also reflected in the store layout. The typical Abercrombie store has a luxury look, is rather dark, has a strong scent, loud music and all employees have a model-look. This makes Abercrombie stand out and attract more customers (Driessen, 2005). However, their strategy might not always be successful with different target groups. Some potential customers could dislike the shop environment and be scared off by the loud music or the heavy fragrance, for example.
A store with a completely different strategy is the Primark. The Primark is known for its cheap clothing, bright lights and its messiness. The messy organization of the stores has even led to the creation of online blogs describing the most effective strategy for visiting a Primark store (e.g. Zoella, 2010). Despite the unattractive shopper experience, the concept of the Primark stores has shown to be an extreme success. This success can be mainly attributed to the low prices levels, which makes customers care less about the messy organization of the stores.
By knowing what triggers certain emotions, brand managers will be better able to optimize the shopping experience. Up till now, most research has looked at the effect of environmental cues on the customers’ emotions. Profound research has already been performed on different environmental cues, such as scent, lighting, and temperature. These research studies have followed the Stimuli Organism Response (SOR) framework, in which the stimuli represent the environmental cues, the organism entail different emotions states (e.g. pleasure and arousal) and the response results in approach or avoidance behavior. Several studies have investigated how these environmental cues could lead to certain positive or negative emotions (Spies, Hesse, & Loesch ,1997; Kumar & Kim, 2014). They found that the in-store environmental cues have a strong relation with accompanying emotions. As stated by Spies, Hesse & Loesch (1997) The mood is improved in a pleasant atmosphere, and deteriorated in a less pleasant atmosphere (p. 13). Baker et al. (2002) specifically researched social, design and ambient cues and how this effect store patronage, by including several emotional aspects. This resulted in the conclusion that shopping in a positive mood is good for sales.
A more thorough understanding of the store environment, and the related emotions, provides management and science with a better understanding of the human mind and the resulting behavior. Currently, research has been performed on store layout, of which merchandise presentation was only a small part. Within this study, the focus will be on the tidy presentation of the merchandise.
As a shop owner, it is important that the shopper spends a lot of time in the store, because it will increase the willingness to spend more (Donovan & Rossiter, 1982). Understanding the different aspects of customer behavior is relevant, because management can discard the negative aspects in the store. In addition, a better understanding of the basic elements in the literature on store environments could shed new light on existing literature.
In general, it is relevant for the managers to be able to understand the state of mind of his customers. Having customers that feel frustrated or stressed might lead to negative word-of-mouth, disloyal customers, and negative product evaluations (d’Astous, 2000). These emotions could harm the business. Optimizing the customer experience, by providing a tidy store, is therefore an important aspect of in-store management. By increasing the understanding of the customer shopping behavior, management will be better able to adjust the store environment to customers’ wishes. This could involve hiring more staff, which will be more expensive, or by better allocating the staff’s resources. Management could achieve this by educating the staff, so they can work on different departments and it can also improve the staff’s overall motivation. Including several store specific attributes into this research will increase the explanatory value of this research. It would be relevant to take a look at whether the shopper is influenced by the price image of the store and by additional price promotions. A shopper might already expect a less tidy environment in stores with a lower price image like H&M or Primark than the, for example, Tommy Hilfiger shop. The expectation of going to the H&M might be different, and the (lack of) tidiness might then be less frustrating. Customers can still expect a tidy store, but because the price is lower, the customer could have different expectations (Gardner & Siomkos, 1985). If there is a seasonal sale in this shop, the customers’ main reason to enter is the price of the merchandise; this could result in even lower attention for the presentation of the merchandise (either messy or tidy).
shop does not fulfill this perception. The Tommy Hilfiger store could be perceived as a high priced store, with only regular price promotions. However, during these promotions, the store could be messier. Customer might get frustrated because it is an expensive brand, but perhaps less frustrated because there is a price promotion.
In general, when a customer enters a store that has a seasonal sale, the main focus for entering this store with a sale is the price (Jin and Kim, 2003). Thus, often the customer is already in a positive mood because of the discounted price. This could influence the overall perception of a store.
Overall, it is therefore relevant to investigate if it is more frustrating for the customer when a store has a high price image, than when a store has a low price image. And whether there is a seasonal sale/price promotion in this store.
This results in the following research questions:
To what extent does (the lack of) shop tidiness lead to shopper frustration? Is this relation strengthened by the price image of the store, and how do price promotions influence this relation?
positive. Thus, providing a better shopping environment will have a positive effect on the evaluation and will increase the stay of the customer in store, and herewith also the likelihood to purchase. Furthermore, this research could contribute to existing literature about brand perception. It is relevant to see if tidiness has a different influence on the shopper when the store image is also different. Koo and Kim (2013) researched the concept of store love in combination with arousal and pleasure. These authors state that the concept of store love is rather broad, hence advice to perform more research on the antecedents of store love, such as store image (p.104).
2. Literature
The current study focuses on the in-store shopping state-of-mind of the potential customer. The in-store shopping literature is part of the main concept of shopper marketing. Shopper marketing is becoming a more important research field, due to the recent changes in technology and business landscape are there new research opportunities that can shed new light on older literature (Bae & Hu, 2010). The concept of shopper marketing is recently clearly defined by Shankar et al. (2011) as the ‘planning and execution of all marketing activities that influence a shopper along –and beyond- the entire path-to-purchase, from the point at which the motivation to shop first emerges through to purchase, consumption, repurchase, and recommendation’ (p.3). As the definition by Shankar et al. states research can focus upon several steps within the path-to-purchase which can be both offline and online. This research will be focusing on the customer while navigating through an offline store.
As mentioned in the introduction, the shopping environment can have different effects on shopper emotions. Therefore, first a deeper understanding of the emotions in shopper marketing theory will be provided, which will, later on, be combined with shopping environmental cues. 2.1. Emotions
Frustration is seen as an emotion that appears if the object or event is negative in its impact on them, and yet not fundamentally negative in character (p.264). The feeling of disgust appears in a situation, when an object or event is bad or ‘rotten’ in character. This means that frustration is caused by a circumstance (see figure 1). Therefore, the negative emotion of frustration clearly fits this research. An untidy shopper environment is caused by a circumstance, such as crowdedness or lack of staff.
This is confirmed by another definition of frustration provided by Scherer (2001) who state that frustration occurs when events are “obstructive for goal attainment, by putting a goal or need satisfaction out of reach, delaying its attainment, or requiring additional effort” (p. 96). This definition clearly states that, when a customer needs to put additional effort to get a product, it is increasingly frustrating. This definition entails the previous definition of Roseman, Antoniou, and Jose, but includes a practical tone, which is relevant for this research.
2.2. Emotions and Environmental cues
Research on store environments has been performed already a few decades. The effect in-store environments have on shopper emotions dates back to the ’70 and ’80 with research by e.g., Kotler (1974) and Donovan and Rossiter (1982). Since that time, more thorough research has been conducted on in-store environmental cues. Table 1 summarizes key prior research studies on in-store environmental cues and store choice criteria.
Variable Authors
Music Morrison et al., (2011); Garlin & Owen,
(2006); Yalch & Spangenberg, (2000); Matilla & Wirtz, (2001); Walsh et al, (2011); Baker et al., (2002)
Scent Morrison et al., (2011); Mattila & Wirtz,
(2001); Walsh et al, (2011);
Design Cue Kusumowidagdo, Sachari, and Widodo (2012); Baker et al., (2002)
Ambient Cue Kumar & Kim, (2014); Walsh et al, (2011); Spies, Hesse & Loesch, (1997)
Merchandise Cue Kumar & Kim, (2014);
Social Cue Kumar & Kim, (2014); Baker et al., (2002)
Service quality perceptions Walsh et al, (2011);
Price perception Walsh et al, (2011);
Table 1 - Previously performed research
As can be seen in the table, much research has been performed on the individual aspects of in-store environments. These previously performed researches provide a good perspective of what has been researched before, and which has not been researched. In addition, it provides knowledge to build this research upon.
Wu et al. (2013) state the importance of visual cues. These authors described the influence visual cues have on customer behavior. Visual cues are often the first way a customer is informed about the shop. This means that when the customer does not like the appearance of the shop, he or she will leave the store. Also Underhill (1999) state that shoppers are influenced or susceptible by the impression the store provides, which will have an impact on the decision making process. The customer might be deciding to stay or to leave, based on whether he or she finds the environment pleasing or not. The decision the customer will make, is highly relevant for the conversion made (Underhill, 1999). This is also found by Lea-Greenwood (1998), who states that the presentation of merchandise is relevant for the shopper, especially to establish a first impression.
which indicates that an appreciated store environment has a positive effect on the shopper’s store behavior (Mehrebian & Russel, 1974; Spies, Hesse and Loesch, 1997). This reveals itself in that the customer has a more positive mood, hence will buy more products (Spies, Hesse, Loesch, 1997).
Research by Turley and Milliman (2000) developed a Stimuli-Organism-Response framework in which they included variables that have been found until 2000. Under the header ‘store layout and design’, the authors included placement of merchandise, and grouping of merchandise, among others. These variables are part of providing the customer with a careful environment (Bitner, 1992). These environmental factors help customers orientate. Therefore, these variables are an important part of the store layout. In addition, the variable cleanliness is placed under the header ‘general interior variables’. Tidiness could be seen as a combination of the two. When the merchandise is placed in an illogical manner or when the merchandise is grouped in an unexpected manner, it could be seen as untidy. However, as a store isn’t clean, it could also be perceived as untidy. This is also reflected in a description of messiness by Doucé et al. (2014). The authors state that messiness is a multi-faceted concept (p.353), hence it consists of several characteristics, namely: dirty, disorganized, complex, cluttered, turbulent, messy, disorderly and untidy (p.354). Besides this description of messiness by Doucé et al., a look will be taken on the more general definition of tidiness: ‘Arranged neatly and in order’ (Oxford Dictionary). Applying this definition into a store environment will give us the following ‘store tidiness’ definition: ‘The in order, and neatly arranged presentation of the merchandise in a store’. Thus, a look will be taken at the presentation of the merchandise. For example, in apparel stores the clothes could be presented properly folded, or could be put unfolded on a big pile of other clothing (is e.g. disorganized, cluttered, and disorderly). The latter could be seen as messy, and will be used to research the relation between frustration and untidiness.
from store environment cues. However, these authors include a large definition of store design perceptions, as they include confusing store layout, finding the merchandise quickly, and easy entering and leaving the store. When a store is messy, it is harder to find the merchandise quickly. This means that the increased difficulty of finding the merchandise will increase the emotional labor or mental stress. This difficulty of finding merchandise increases emotional labor, however can also be perceived as frustrating. As is also mentioned by Scherer (2001); when it is necessary to put extra effort to get a product, it becomes more frustrating. Thus, the difficulty of finding merchandise leads to a higher degree of frustration.
This leads to the following hypothesis:
H1: The messiness/untidiness of a store is positively related to frustration. 2.3. Price strategies
Each store has a certain price-level image. Price image is defined by Hamilton and Chernev (2013) as ‘the categorical impression of the aggregate price level of a retailer’ (p. 2). There are different price-level images. Certain stores (often supermarkets) are known for their Every Day Low Pricing (EDLP), and others are known for, the pricing strategy on the other end of the continuum, its High Low (HiLo) strategy (Tang, Bell, and Ho, 2001). Stores with an EDLP strategy are often discounters. They do not put products into additional discounts or promotions. The stores with a HiLo strategy have in general relatively higher prices, but focus on sales and promotions of certain products. According to Baker et al. (2002), the price perception has an influence on the behavioral intentions of the shopper, such as store patronage.
Koo and Kim (2013) state that a well-designed store positively impacts consumer judgment about the merchandise quality in a store. Thus, for an untidy store the customer could perceive the store as having a lower price image, since the overall store quality is also perceived as lower. The untidiness does not correspond to the price image. In a store with a high price image, the customer pays a premium for the service, and quality of the merchandise. Koo and Kim have shown that there is a direct relationship between merchandise presentation and price image. However, it is also possible to incorporate price image from a different angle. As there is a relation between tidiness and price image, it could also be researched whether price image strengthens the relation between tidiness and frustration. When the price image in a store is not able to fulfill the quality demand, the customer might feel that the price image is not corresponding to the customers’ expectations. This discrepancy is likely to increase the frustration of the customer.
Thus, when a customer visits a store with a high price image, they might be more bothered by the mess, as it does not fulfill the expectation. The price image perception makes customers assume that the store has reasons to be more expensive. This could be based on several in-store cues, or services. When the quality demand is satisfied, because the store is tidy, the customer has no reason to become frustrated. But when the price image perception is not met, because the store is messy, the customer becomes more dissatisfied, hence more frustrated. Therefore, the following hypothesis can be derived:
H2: The price image of the store has a positive effect on the relationship between untidiness and consumer frustration
2.4. Discounts and Promotions
store, could also apply for entering a normal store, which has temporal discounts. The main focus of the shopper lies in finding a discount, which means that other in-store cues are likely to be processed only limited.
In apparel stores, promotions are also an extra method to sell the previous (seasonal) collection. Due to the lower prices more people enter the store, this increase of crowding could effect the perception of the environment (Schmidt and Keating, 1979). The customers could perceive the circumstance as less pleasant (Eroglu and Harrel, 1986). The customers might be more frustrated because of the chaos in the store, but on the other hand become more euphoric because of the low prices.
H3: Sales promotions have a moderating effect on the moderation of price image on the relationship between untidiness and consumer frustration (moderated moderation)
2.5. Control variables
Next to the hypotheses it is important to include variables that might influence the proposed relations. When these variables are not included, causal relations could be missed which could cause wrongful conclusions (Leighter and Inoue, 2012). This is also known as omitted variable bias. Therefore, the following control variables will be included in the analysis.
2.5.1. Mood
The dependent variable in this research is frustration. It is important to include the mood of the shopper in the analysis. The initial mood of the shopper could influence the perception he or she has of the situation (Greenland & McGoldrick, 1994). Including this in the model makes measuring the dependent variable more reliable.
2.5.2. Sales Proneness
2.6. Conceptual Model
The expected relations are presented in the following conceptual model (Figure 2):
Store Messiness Store Price Image Frustration In-store Promotions
Figure 2 - Conceptual Model
Control Variables:
3. Data
In this chapter the method for data collection and means for data analysis will be assessed. The hypotheses stated in chapter 2, will be examined in an experimental study. This study will be performed by distributing an online questionnaire.
3.1 Data collection
This study follows a 2x2x2 full factorial design. In the following table, the variables and corresponding values will be presented. A full factorial design means that the each of the possible combinations will be tested in the experiment. This method
Variable Content Condition
Dependent variable Frustration 1) Numerical scale
IV1 (Direct effect) Tidiness 1) Store is tidy
2) Store is untidy
IV2 (Moderator) Price image 1) High price image
2) Low price image IV3
(Moderated Moderator)
In-store sales 1) Sales period 2) No sales period Table 2 - Variables
products are often high-involvement products. This means that the decision to buy is often more deeply considered. During this experiment, the customer should feel more associated with the product, and willing to buy it instantly.
The survey will be based on questions found the Handbook of Marketing Scales and Marketing Scales Handbook. These two books provide an overview of scales that has been used in previously performed research. It makes sure that proper questions are asked and have a known relation.
To make sure that the research is performed properly, there will be one questionnaire in which the 8 different situations will be rotated. This means that some people answer questions about an untidy store, with a low price image, and a promotion period, while others will receive questions of another situation. The different control groups make it possible to compare all variables in the research. Visuals of the different situation will be provided, since this limits the bias in the research, because respondents don’t have to answer from memory. Pictures of a tidy or untidy store visualize the situation best. Besides the pictures an accompanying text will be written, which explains the additional variables, such as price image and price promotion. In Appendix A, the complete questionnaire can be found, including the 8 different scenarios.
3.1.2. Obtaining respondents
To perform a proper research, 30 respondents are necessary per condition. That means that for this study, a full factorial design, a minimum number of 240 respondents (8x30) is necessary. The best way to obtain this number of respondents is by using social media. It is possible to spread the online questionnaire via Facebook and/or Twitter. Friends and family will be able to share the message to their friends. This is an example of snowballing, which would increase the sample size strongly (Polkinghorne, 2005).
3.2. Recoding
analysis, the responses should be recoded. This means that the order of the answer possibilities should be reversed.
3.3. Descriptive
The survey has been online for almost two weeks. During these two weeks, 288 people have opened the survey. Of these 288 people, 80% has completed the survey as a whole. The other incomplete recorded responses are removed from the dataset. This leaves a total dataset of 238 respondents. It is relevant to see how the dataset looks like, an evenly distribution of the demographics over the experimental groups, increases the reliability of the dataset.
In the following graphs, the different experimental groups are stated. The abbreviations are: HP/LP = High price image / Low price image
T / M = Tidy / Messy S / NS = Sales / No Sale
The respondents are distributed over the group as follows:
HP_T_NS HP_M_NS HP_T_S HP_M_S LP_T_NS LP_M_NS LP_T_S LP_M_S Number of
Respondents
29 31 30 28 29 32 30 29
Table 3 - Number of respondents per group
Unfortunately, the groups are not evenly distributed, due to missing values. However each group consists of almost 30 respondents, which means there are enough respondents in each group to perform an analysis.
The following tables describe how the sample looks like. 3.3.1. Gender
Graph 1 - Percentage male/female per group 3.3.2. Age
In graph 2, the average age per group is shown. Besides the age, the standard deviation/range is indicated. The graph shows that in each group the average age is relatively equal, and there is limited fluctuation in the range. This means that the respondents are equally distributed over the groups. In addition, the average age is around 25 and 26. This implies that the average age is relatively low. This is due to the high number of students to which this survey is distributed.
Graph 2 – Average Age per group 3.3.3. Education
In the next graph (3), the education level per group is presented. As can be seen is the average level high. Most respondents received or follow either a Bachelor degree or a Masters degree. This is also due to the large number of respondents in the sample. In each group, there are only a limited number of people following the lowest degree of education, which is not corresponding to the Dutch public.
Graph 3 - Education level per group 3.3.4. Residence
In the last graph, the place of residence is shown. Most respondents reside in Groningen. Besides Groningen, live most people in the western part of the Netherlands, namely Noord-Holland en Zuid-Holland. Besides Dutch respondents, did foreigners also frequently fill in the survey.
Unfortunately, no residents from Zeeland en Flevoland have filled in the questionnaire.
The previous graphs indicate a rather even distribution of the demographics among the groups. However, it is important to check whether there is no significant difference between the groups. If this is the case, the results could be biased towards one specific demographic group. The different demographical variables could have an influence on the dependent variable (Leliveld & Wiebenga, 2014), which could bias the results. When this bias is found among the groups, the demographic characteristic should be controlled for. The bias can be statistically tested by means of a t-test. In this research all demographic characteristics have been tested, but no significant results are found, hence for no demographic characteristic should be controlled for. This can be seen in the following table:
F-value Significance
Gender 2,248 ,135
Age ,952 ,330
Education ,882 ,349
Province ,068 ,795
4. Methodology
In this chapter, the method of analysis will be presented. First, a factor analysis will be presented for data reduction. In continuation of this, validation techniques for a regression analysis will be presented. Finally the methods for performing a regression analysis will be shown.
4.1. Factor analysis
By means of factor analysis, internal consistency can be found among the independent variables. This makes it possible to reduce the data and summarize the data. (Malthotra, 2007). The factors retrieved form factor analysis can be used in the ANOVA.
An additional advantage of a factor analysis is that it limits the risk of multicollinearity. When all the individual variables with a high internal consistency will be included in an interaction effect, it will result in a high degree of correlation among the variables, hence multicollinearity will be found (Leeflang, et al. 2014). As will be explained later, multicollinearity could have a bias the results from the analysis. With factor analysis, this effect is limited, since several variables that have a high degree of correlation are combined into variable. Furthermore, it is a way to examine the overlapping variation of the variables. (Leeflang et al., 2014, p.114).
4.1.1. Validity
4.1.2. Factor extraction method
The communalities can be extracted by several different extraction methods. In this case, the principle components analysis is used. This method considers the complete variation in the dataset (Malhotra, 2007).
4.1.3. Determining the factors
There are several ways of determining the number of factors. First, the number of factors can be estimated with prior knowledge. Based on knowledge about the research an a priori determination of factors can be performed. In addition, it can be based on the values provided in the table “Total Variance Explained”, which are the “Eigenvalues”. The number of factors is then chosen, when the Eigenvalue has a value higher than 1. When the number is less than 1, the variable does not gain explanatory value when transformed to a factor. By means of a scree-plot, a visualization of the number of factors is provided. The point where the scree in the graph starts, visualizes the number of factors.
4.1.4. Assigning the factors
Based on the number of factors chosen before, the component matrix provides the component scores per variable per factor. Each variable scoring higher than 0.6 is included in one factor. 4.2. ANOVA
With the factors established from factor analysis, an ANOVA can be performed. An ANOVA stands for ‘Analysis of variance’. An ANOVA determines the significance level of the difference between more than two or more means, as well as to find out what he effect of one or more nominal variables is (Janssen et al. 2012, p.72). After the ANOVA, a regression analysis will be performed, to clearly distinguish the relations among the variables.
4.3 Multiple regression
After the ANOVA, a multiple regression will be performed. A regression analysis has been defined by Malhotra (2007) as a ‘statistical procedure for analyzing associative relationship
between a metric dependent variable and one or more variables” (p.542). The regression
variables. In this research, the dependent variable entails the degree of frustration, and the independent variables include the degree of tidiness, price image, and sales.
4.3.1. Validation for ANOVA and Regression
When performing an ANOVA or a regression analysis, it is important to take several validation criteria into account. When these criteria are not met, the ANOVA and the regression analysis estimates cannot be seen as valid. The following validation criteria are for the regression analysis and/or for the ANOVA.
4.3.1.1. R-square
The R-square measure describes the percentage of variation in the model. A high R-square measure indicates a higher model fit, a lower R-square value means a lower model fit. This is due to a lower variance of the explanatory variables (Leeflang et al., 2014). A low R-square represents a model with less explanatory value, therefore it is necessary to strive for a model with a high R-square value.
4.3.1.2. Levene’s Test of Equality of Error Variances (ANOVA)
The Levene’s Test for Equality of Error Variances tests the error variance in the groups. For a good measurement is it important that the different experimental group have equal error variances. The following hypotheses must be tested:
H0 = The error variances in the groups are equal H1 = The error variances in the groups are different
When the H0 is rejected, it is important to remedy this, which can be done by transforming the dependent variable or applying a non-parametric analysis (Janssens et al. 2012).
4.3.1.3. Multicollinearity (Regression)
the VIF should not exceed ten, and the tolerance level should be lower than 0,1. When there is multicollinearity is can be resolved in several different ways (Leeflang 2014, p.114):
1) Obtain more data that is relevant for the problem, 2) Reformulate the model
3) Create new predictors
4) Apply estimation methods specifically for multicollinearity 5) Eliminate a predictor variable with a t-ratio close to zero
4.3.1.4. Residual diagnostics
The validation criteria can help identify outliers from the model. Outliers can be identified when plotting a box-plot. When there are individual data-points, which are not included in the boxplot, these could be seen as outliers. These outliers could influence the results, and should therefore be removed.
4.3.1.5. Non-Normality
Another important assumption that should be checked is whether the disturbances are normally distributed. This assumption can be checked with the Kolmogorov-Smirnov and Shapiro-Wilk tests. They assume two hypotheses:
H0 = The disturbances are normally distributed H1 = The disturbances are not normally distributed.
The disturbances should be normally distributed for the standard hypothesis testing and the confidence intervals to be applicable (Leeflang et al., 2014). Non-normality can be solved by applying the logs to the variables and by removing outliers that can disturb the normality plot.
4.3.1.6. Heteroscedasticity (Regression)
be detected by finding patterns in the residual plots, and by performing a Goldfeld-Quandt test. The Goldfeld-Quandt test will be performed when the residual plot is inconclusive:
With:
SSR = the sum of the squared residuals O = Observations
P = Parameters.
By means of a F-distribution, the significance level can be tested: H0 = The error-term is homoscedastic
H1 = The error-term is heteroscedastic
5. Results
5.1. Control variablesIn the following tables, the results from the factor analyses are presented. For all analyses the Keiser-Mayer-Oklin test resulted in a score above 0,6 and the Bartlett’s test for sphericity was found significant, this means that it is appropriate to continue with factor analysis.
The first variable set entailed variables regarding the initial mood of the respondent:
Variable Factor: Initial_Mood
Currently I am in a good mood.
,719 As I answer these questions
I feel cheerful.
,635 For some reason I feel not
very comfortable right now.
,664 At this moment I feel edge
or irritable.
,624
KMO ,615
Bartlett ,000
Table 5 - Factor 1 - Initial Mood
The second analysis is performed on the scale ‘sales proneness’ of the respondent:
Variable Factor: Sales_proneness1 Factor 2: Sales-Proneness2
If a product is on sale, that can be a reason for me to buy it
,587 ,635
When I buy a brand that’s on sale, I feel I am getting a good deal.
,519 ,708
I have favorite brands, but most of the times I buy the brand that’s on sale.
,752 -,352
I am more likely to buy brands that are on sale.
,829 -,293
Compared to most people, I am more likely to buy
brands that are on sale.
KMO ,696
Bartlett ,000
Table 6 - Factor 2 & 3 Sales Proneness
In table 6, it is not possible to include the questions in one factor. There is a distinction between two types on sales proneness. It could be argued that Factor 2 entails sales with regard to one specific brand, while Factor 3 entails sale behavior with regard to buying a brand without any specific preference. Sales Proneness2 fits best in this analysis, since it regards the overall sales perception of the consumers. Therefore, this factor will be included in the analysis.
5.2. ANOVA
As mentioned before, for testing the hypotheses an ANOVA test will be used. Within the ANOVA the dependent variable Frustration and the independent variables Tidiness, Price Image and Sales will be incorporated. Besides these independent variables, covariates will be incorporated. Since the dependent variable is Frustration, the initial mood could influence the final state-of-mind of the respondent. Therefore, the covariate factor ‘initial mood’ will be included in the analysis. In addition, could the effect of a sale or not, be influenced by how prone respondents are for sales. It is therefore important that the sales proneness will be controlled for in the analysis. Thus, the second covariate will be the two factors ‘Sales Proneness’.
The following table provides the results from the ANOVA.
Sales
No Sales
High Price Image Low Price Image
Tidy Store 5.282 4.745 5.375 4.628 Messy Store 3.982 3.038 3.938 3.736 Table 7 - ANOVA Means per group
regression analysis will be performed to clearly understand the relations in the analysis, as hypothesized in chapter 2. ANOVA test Intercept 3751,526 *** Tidiness 86,719 *** Price Image 1,158
Tidiness * Price Image (Moderator) 1,453 Tidiness * Price Image * Sales (Moderated
Moderator)
1,318
Price Image * Sales ,936
Sales 17.757 ***
R-square ,331
R-square ,311
Levene’s Test of Equality of Error Variances
,236
Significance * = p<,1; ** = p<,05; *** = p<,01
Table 8 - Results ANOVA 5.2.1. Validation
Before the results will be interpreted, the validation assumptions will be discussed. The first assumption checked is whether the overall model is significant. The ANOVA output showed a p-value of ,000, which indicates that the corrected model is significant. This means that the first assumption is met.
penalty for adding variables, however this is penalty is not too large, that it will influence the results.
The third assumption relates to the Levene’s Test of Equality of Error Variances. As mentioned in paragraph 4.3.1.2., the assumption regarding Levene’s Test should be insignificant. Fortunately shows the output a value of ,236, so the alternative hypothesis is not satisfied. The assumptions regarding non-normality and outliers will be discussed after the regression analysis, since these results will be similar.
5.3. Regression
34 Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model
35 5.3.1. Model selection
5.3.1.1 R-square
In the model, the r-square and the r-square adjusted are provided. As mentioned before, the R-square provides the explanatory value of the model. The estimates range from ,312 in the simple model, to ,342 in the full model. However, the R-square adjusted provides a better estimate, as it is adjusts for the added parameters in the model. The model with the highest R-square adjusted, can be seen as the model with the highest explanatory value, which is in model 31.8%. When basing the model selection on the square measures, Model 7 includes almost the highest R-square and the highest value for the R-R-square adjusted.
5.3.1.2. Information Criteria
Besides basing the model selection on the R-square, there are also Information Criteria on which model selection could be based upon. The first Information Criteria is the Akaikies Information Criteria (AIC). This includes the number of parameters (npar), the sample size (n), and the sum of squared residuals (SSR):
AIC = n ln(SSE) − n ln(n) + 2p
Besides the AIC, is the Bayesian Information Criteria (BIC). The BIC selects models based on the following formula:
BIC = n ln(SSE) − n ln(n) + ln(n)p
For both the AIC and the BIC, is the model with the lowest value the best fit. The result can be compared in the following table:
36 N 238 238 238 238 238 238 238 Npar 3 4 4 4 6 7 8 SSR 290,052 288,606 289,577 289,656 287,616 282,861 277,990 AIC 53,074 53,884 54,684 54,749 57,067 55,099 52,965 BIC 2557,73 3861,32 3860,52 3860,45 6466,94 7773,30 9079,84
Table 10 – Results AIC and BIC
As can be seen in the tables, the AIC provides the lowest number for Model 7. This means that Model 7 is the best choice. Unfortunately, the BIC is inconclusive.
Following the criteria for model selection, as mentioned above, the best model would be Model 7.
5.3.2. Validation
As mentioned in chapter 4, there are several assumptions, which must be satisfied before interpreting the output. As the previous paragraphs explained, will there be continued with the estimations of Model 7.
5.3.2.1. R-square
The R-square is besides a model selection criterion, also a validation method. As mentioned before, the explanatory value for Model 7 is 34.1%, when penalizing for additional variables, the R-square adjusted reveals an explanatory value of 31,8%.
5.3.2.2. Multicollinearity
The next assumption that should be satisfied is that there should not be any correlation among the variables. Table 11 shows the Variance Inflation Factor and the Tolerance level. As mentioned in chapter 4, the Tolerance should be between 0,1 and 1, and the VIF should not exceed 10. The output stated in table 11 indicates no extreme multicollinearity. The VIF has
Model 8 Model 9 Model 10 Model 11 Model 12
N 238 238 238 238 238
Npar 8 9 10 10 11
increased to 7 for the three-way interaction, however not severe enough to state that the assumption for multicollinearity is not met (Leeflang, et al., 2014).
Tolerance VIF Tidiness ,255 3,915 Price Image ,264 3,783 Sales ,255 3,920 Tidiness*Price Image ,170 5,889 Tidiness*Price Image* Sales ,141 7,092 Tidiness*Sales ,167 5,995
Price Image * Sales ,170 5,889
Sale Proneness ,983 1,018
Table 11 – Multicollinearity
5.3.2.3. Residual Statistics
The residual statistics indicate whether there are any outliers in the data. This is shown in the following box-plot of the residuals. As can be seen covers the box-plot all data-points. This means that there are no outliers that could bias the results.
38
5.3.2.4. Non-normality
With the Kolmogorov-Smirnov and the Shapiro-Wilkinson test statistic, the normality of the data is tested. A significant value for both the Kolmogorov-Smirnov and the Shapiro-Wilk means that the data is not normally distributed. As can be seen in table 12, the test are not found to be significant, hence the data is normally distributed. This assumption for validation has also been met.
5.3.2.5. Heteroscedasticity
The last assumption to be met is with regard of patterns in the residuals. A pattern in the residuals assumes that the residuals are related. However, it is important that the residuals are unbiased. This can be tested visually. The residuals are plotted against the predicted value. In figure 5, the scatterplot is presented. The plot shows no pattern in the residuals, which indicates that there is no heteroscedasticity in the data. An example of heteroscedastic plots of residuals can be found in Appendix B. The last assumption is also met.
Figure 4 - Residuals plot
5.3. Interpretation of the results
For the ease of interpretation, the results up to model 7 are presented in the next table: Table 12 - Non-normality tests
Kolmogorov – Smirnov Shapiro – Wilk
Statistic Df Sig. Statistic Df Sig.
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Intercept 3,440*** (,142) 3,516*** (,158) 3,482*** (,157) 3,479*** (,158) 3,558*** (,173) 3,750*** (,196) 3,736** * (,196) Tidiness 1,349*** (,144) 1,194*** (,203) 1,261*** (,203) 1,309*** (,161) 1,063*** (,270) ,871** (,284) ,887*** (,283) Price Image -,140 (,144) -,295 (,203) -,140 (,145) -,181 (,162) -,295 (,204) -,685* (,279) -,702** (,278) Sales ,571*** (,144) ,571*** (,144) ,482* (,204) ,530*** (,161) ,482* (,204) ,078 (,284) ,106 (,283) Tidy * PI ,312 (,289) ,398 (,357) ,789** (,404) ,823** (,403) Tidy * Sales ,179 (,289) ,264 (,355) ,668* (,405) ,652* (,403) Tidy*Price* Sales ,163 (,289) -,170 (,411) -,992* (,578) -1,068* (,573) PI*Sales ,822** (,406) ,842** (,403) Sales Proneness ,145** (,072) Mood Tidy*Mood Tidy*Prone ness R2 ,312 ,316 ,313 ,313 ,317 ,329 ,341 R2 adjusted ,303 ,304 ,302 ,301 ,300 ,309 ,318
Table 13 - Output for Model 7
40 significantly strengthening effect between price image and sales, which implies that there is a positive relation between sales and price image. Finally, the covariate Sales Proneness has a significant effect on the relations presented in the model. Sales proneness controls for the how sale sensitive the consumers are, and prevents an omitted variable bias.
Hypothesis Supported/
Not supported
H1: The messiness/untidiness of a store is positively related to frustration.
✔***
H2: The price image of the store has a positive relation on the effect between
untidiness and frustration
✔**
H3: Sales promotions have a moderating relation on the moderation between
price image and the direct effect between untidiness and frustration
✔*
***= fully supported, **=supported, *= marginally supported6. Discussion & Conclusions
In the discussion and conclusion, the answer will be provided on the research questions. This will be based on the hypothesis stated in chapter 2, and the results from chapter 5.
The research questions stated in the introduction are:
To what extent does (the lack of) shop tidiness lead to shopper frustration? Is this relation strengthened by the store image, and how do price promotions influence this relation?
6.1 Discussion
As mentioned in the results section, all hypotheses are supported. However, the third hypothesis is only marginally supported.
The results indicate that the store tidiness has a strong effect on the degree of frustration. The customers clearly have a strong negative attitude towards untidy stores. Research by Kerfoot, Davies & Ward (2003) supports this finding, as the authors state that the presentation of merchandise is highly relevant in a store environment. This research extended on the current general vision of merchandise presentation. While the researchers mainly focus on the merchandise presentation or store environment, it is seen that the tidiness of stores is a relevant part of these literature streams. The more specific understanding of the store environment increases the understanding of the shopper’s behavior.
42 account when entering stores (Turley and Milliman, 2000). This research shows that story tidiness is also a relevant factor in evaluating store environments. Furthermore, with a perceived higher quality level, often the light in the store is friendlier, and less bright, the music is more calming and classical (Gardner and Siomkos 1985), now it can be seen that the tidiness is also an aspect that is important for stores with a high price image.
The last inclusion into the model is the sales period. The results indicate that when a store has a sale, the customers become less frustrated when a low price image store is untidy. This means that the customers are mainly focusing on the price during a sale. As mentioned in chapter 2, the customers might be in a relatively good mood in the shop, as a result of the attractive price promotions. This distracts them from the store environment. The customers are less bothered by the presentation of the store, since the store is providing them a lower price. In addition, it is possible that the customer has a different expectation in advance when there is a sale. The store is probably more crowded, at least at the places where the clothing is on sale. The customer is often aware of this phenomenon before entering the store or heading to the table. Since the customer anticipates on the situation, they might be less bothered by the untidiness of the clothing.
6.2. Conclusions
Concluding, tidiness has a strong positive relation with frustration. The tidier the store is, the less frustrated the shopper becomes, which is good for the store performance in terms of expected sales and reputation. In addition, the price level has a significant influence on this relation. The more expensive the store is, but still untidy, the shopper feels increasingly frustrated. Thus, for customers untidy stores are highly frustrating, and even more when it is an expensive shop. Furthermore, when shops have an e.g. summer sale, the customers are less inclined to feel frustrated when the shop is not tidy and less expensive.
6.3. Limitations & direction for further research
population, and could perhaps respond different to situations. In addition, the sample is also relatively young. The average age is 26 years old. The generalizability of the outcomes could therefore be limited. For further research, it is relevant to try to have a sample, which is representable for the whole society.
Another limitation could be the specificity of the brand store chosen. The H&M and Tommy Hilfiger store are both well-known store, however these specific stores could also influence the perception of the respondents. Store lovers could respond differently to the questions in the survey than store haters. Other researchers could include different stores in the analysis, more different types of stores, or do not use brand names in the research at all. Furthermore, the researchers could control for store lovers.
A third limitation could be that tidiness is only measured by means of folded or unfolded sweaters. Existing research by Kerfoot, Davies & Ward (2003), indicate that people could respond differently to clothing which is folded or hanging in the shelves. In future research the researchers could investigate different ways of tidiness, and which presentation method is perceived as untidy.
35
7. Managerial Implications
This chapter describes the importance of the results for management and how they can implement this in their daily business.
From the results, it could be seen that tidiness, regardless of price image or sales, has a strong effect on frustration. This means for management, that the store environment should always be tidy. The customer perceives an untidy store environment as highly frustrating. Theory on approach and avoidance behavior shows that the negative emotion could lead to avoidance behavior (Mehrebian & Russel, 1974). When the customer perceives the store an untidy, they might leave the store, or worst case, do not come back to the store at all.
In addition, the customer might tell his/her friends about its experience with the store, hence negative word-of-mouth is created (d’Astous, 2000). Therefore, it is important that the customer experience is always positive; hence the store should be tidy.
When the store has a higher price image, the effect is even stronger. When the customer visits a more expensive store he or she expects the store to be tidy, as quality and price are closely related. The management of high priced stores should therefore, focus even more on store tidiness. Since the effect for higher priced stores is stronger. This means that the customers are in a more negative mood, which could lead to potential avoidance behavior (Bitner, 1992).
38 However, hiring additional staff is highly expensive. It is therefore, more important that the current staff is well trained and managed. The main antecedent for an improved store environment is related to management. According to Van Schaik (2012) it is important that the staff is properly managed, hence the store will perform more profitable. Part of being well managed is that the staff is also properly trained. When the staff is not able to anticipate on the chaotic situation, more staff will not increase the tidiness in the store. The staff should be well trained, to be capable of adjusting to the store situation. In addition, the staff would become more motivated when management invests in more training. Therefore, training and guidance by store management is key (Van Schaik, 2012).
During sales periods, the customers are slightly less frustrated; therefore the store could be untidier. However, this should only reflect on the clothing, which is in sale. The customers that could bother less on untidily folded clothing, which is on sale, rather than clothing that is not on sale. Thus, it is important to have the rest of the store properly ordered. In addition, it is important that the overall store is not seen as untidy. This distinction should be clear for management.
References
Abercrombie & Fitch Corporate website (2012) Abercrombie & Fitch Brand Protection (Accessed on 18th of November 2014) [Available at:
http://eu.abercrombie.com/webapp/wcs/stores/servlet/CustomerService?textKey=HELP_BRAN DPROTECTION&catalogId=11556&langId=-1&pageName=brand-protection&storeId=19159] d’Astous, Alain (2000). Irritating aspect of the shopper environment. Journal of Business Research , 49, 149-156.
Babin, Barry J., David M. Hardsty, and Tracy A. Suter (2003) Color and shopping intentions: The intervening effect of price fairness and perceived affect. Journal of Business Research 56
(7), 541-551.
Bae, Sang Hee, and Mantian Hu. (2010) Shopper marketing. Conference Summary from Marketing Science Institute Conference, New York University. 14-15 June 2010, New York Bagozzi, Richard P., Mahesh Gopinath, and Prashanth U. Nyer, (1999) The role of emotions in marketing, Journal of the Academy of Marketing Science 27 (2), 184-206.
Baker, Julie, Dhruv Grewal, and A. Parasuraman (1994). The influence of store environment on quality inferences and store image. Journal of Academic Marketing Science (22), 328-339. Baker, Julie, A. Parasuraman, Dhruv Grewal, and Glenn B. Voss (2002) The influence of multiple store environmental cues on perceived merchandise value and patronage intentions.
Journal of Marketing 66, 120-141.
Baumgarten, Steven A., and James S. Hensel (1987) Enhancing the perceived quality of medical service delivery systems. In: Add value to your service. Ed: Carol Suprenant, Chicago Il; American Marketing Association, 105-110.
Bitner, Mary Jo (1992) Servicescapes: The impact of physical surroundings on customers and employees, Journal of Marketing 56 (April), 57–71.
Brown, F.E. (1969), Price Image Versus Price Reality, Journal of Marketing Research (6), 185-191.
40 Dodds, William B., Kent B. Monroe, and Dhruv Grewal (1991) Effects of Price, Brand, and Store Information on Buyers ' Product Evaluations. Journal of Marketing Research 28 (August), 307-319.
Doucé, Lieve, Wim Janssens, Gilbert Swinnen, Koenraad Van Cleempoel (2014) Influencing consumer reactions towards a tidy versus a messy store using pleasant ambient scents, Journal of
Environmental Psychology 40, 351-358.
Driessen, Claire E. (2005) Message Communication in advertising: Selling the Abercrombie and Fitch image, UW-L Journal of Undergraduate Research VIII, 1-12.
Eroglu, Sevgin, and Gilbert D. Harrell (1986) Retail crowding: Theoretical and strategic implications, Journal of Retailing 62 (4), 346-363
Ganesh, Jaishankar, Kristy E. Reynolds, Michael Luckett, and Nadia Pomirleanu (2010) Online shopper motivations, and e-store attributes: An examination of online patronage behavior and shopper typologies, Journal of Retailing 86 (1), 106-115.
Gardner, Meryl P. and George J. Siomkos (1985) Toward a methodology for assessing affects of in-store atmosphere, In Advances in Consumer Research, Ed, Richard Lutz, Chigago IL; Association for Consumer Research, 27-31.
Garlin, F.V. & Owen, K. (2006) Setting the tone with the tune: A meta-analytic review of the effects of background music in retail settings. Journal of Business Research 56 (6), 755-764. Greenland, S.J. & McGoldrick P.J. (1994) Atmospherics, Attitudes and Behaviour: Modelling the Impact of Designed Space, The International Review of Retail, Distribution and Consumer Research (4), 1-16.
Grewal, Dhruv, R. Krishnan, Julie Baker, and Norm Borin, (1998) The effects of store name, brand name, and price discounts on consumers’ evaluations and purchase intentions, Journal of
Retailing 87 (3), 331-353.
Hamilton, Ryan, and Alexander Chernev (2013) Low prices are just the beginning: Price image in retail management, Journal of Marketing 77, 1-20.
Jin, Byoungho, and Jai-Ok Kim (2003) A typology of Korean discount shoppers: shopping motives, store attributes and outcomes. International Journal of Service Industry Management
14 (4), 396-419.
Kaltcheva, Velitchka .D. and Barton A. Weitz (2006) When should a retailer create an exciting store environment? Journal of Marketing 70, 107-118.
Kerfoot, Shona, Barry Davies, and Philippa Ward (2003) Visual merchandising and the creation of discernible retail brands, International Journal of Retail & Distribution Management 31 (3), 143-152.
Koo, Wanmo, & Youn-Kyung Kim (2013) Impacts of store environmental cues on store love and loyalty: Single-brand apparel retailers. Journal of International Consumer Marketing (25), 94-106.
Kotler, Philip (1974) Atmospherics as a marketing tool, Journal of Retailing 49 (4), 48-64. Kumar, Archana and Youn-Kyung Kim (2014) The store-as-a-brand strategy: The effect of store environment on customer responses. Journal of Retailing and Consumer Services (21), 685-695. Kusumowidagdo, Astrid, Agus Sachari, and Pribadi Widodo (2012) Shoppers’ perception on physical condition of shopping centers’ atmosphere at different lifecycle, International Journal
of Business Studies, 87-100.
Lea-Greenwood, Gaynor (1998) Visual merchandising: a neglected area in UK fashion retailing?
International Journal of Retailing and Distribution Management 18 (4), 21-31.
Leeflang, Peter S.H., Tammo .H.A Bijmolt, Koen Pauwels, and Jaap E. Wieringa (2014) Modeling Markets, Springer
Leightner, Jonathan, E., & Inoue, Tomoo (2012) Analysis using the relative vertical postition of observations, Advances in Decision Sciences 2012, 1-25
Lichtenstein, Donald R., Nancy M. Ridgway, and Richard G. Netemeyer, R.G. (1993) Price perception and consumer shopping behavior: A field study, Journal of Marketing Research 30
(2), 234-245.
42 Matilla, Anna S. & Jochen Wirtz, (2001) Congruency of scent and music as a driver of in-store evaluations and behavior. Journal of Retailing 77 (2), 273-289.
Mehrebian, Albert, and James A. Russel (1974) An approach to environmental psychology, MIT
Press, Cambridge, MA.
Mizerski, Richard W. (1982) An Attribution Explanation of the Disproportionate Influence of Unfavorable Information, Journal of Consumer Research 9 (December), 301–310
Morrison, Michael, Sarah Gan, Chris Dubelaar, Harmen Oppewal (2011) In-store music and aroma influences on shopper behavior and satisfaction, Retailing Evolution Research 64 (6) 558-564.
O'Neill, Regina M., and David R. Lambert (2001). Exploring the emotional aspects of price.
Psychology & Marketing, 18(3), 217-237
Park, Nam-Kyu, and Cheryl A. Farr (2007) The effects of lighting on Consumers’ emotions and behavioral intentions in a retail environment: A cross-cultural comparison. Journal of Interior
Design 33 (1), 17-31.
Polkinghorne, Donald E. (2005) Language and Meaning: Data collection in qualitative research,
Journal of Counseling Psychology 52 (2), 137-145.
Provoost, Ellen (2013) ‘Dikkerds niet welkom in winkels van Abercrombie & Fitch’ [accessed at
27 October 2014) [Available via:
http://www.hln.be/hln/nl/40/Style/article/detail/1629312/2013/05/08/Dikkerds-niet-welkom-in-winkels-van-Abercrombie-Fitch.dhtml]
Roseman, Ira J. (1984) Cognitive determinants of emotions: A structural theory. In P. Shaver (Ed.), Review of personality and social psychology 5, 11-36 Beverly Hills, CA: Sage
Roseman, Ira J., Ann Aliki Antoniou, and Paul E. Jose (1996) Appraisal determinants of emotions: Constructing a more accurate and comprehensive theory. Cognition and Emotion 10
(3), 241-277.
Scherer, Klaus R. (2001) Appraisal considered as a process of multilevel sequential checking. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processes in emotion, 92–120 New York: Oxford University Press.
Shankar, Venkatesh, J. Jeffrey Inman, Murali Mantrala, Eileen Kelly, and Ross Rizley (2011) Innovations in shopper marketing: Current insights and future research issues. Journal of
Retailing (87S), S29-S42.
Singh, Renu (2006) An empirical investigation into the effects of shopping motivation on store environment-value relationship. Doctoral Dissertation , 1-101.
Spies, Kordelia, Friedrich Hesse, and Kerstin Loesch, (1997) Store atmosphere, mood and purchasing power. International Journal of Research in Marketing (14), 1-17.
Summers, Teresa A. and Paulette R. Herbert (2001) Shedding some light on store atmospherics: Influence of illumination on consumer behavior. Journal of Business Research 54 (2), 145-150 Tang, Christopher S., David R. Bell, and Teck-Hua Ho (2001) Store choice behavior and shopping behavior: How price format works, California Management Review 43 (2), 56-74 Turley, L.W. and Ronald E. Milliman (2000) Atmospheric effects on shopping behavior: a review of the experimental evidence, Journal of Business Research 49 (2), 193-211.
Underhill, Paco. (1999) Why we buy: The science of shopping. Simon & Schuster, New York, NY.
Unruh-Enos, T. (2014) How online retail is quickly taking over the market (accessed at 22 September 2014), [Available at: http://www.smartasset.com/blog/economics-of/how-online-retail-is-quickly-taking-over-the-market/]
Vieira, Valter Alfonso (2013) Stimuli-Organism-Reponse Framework: A meta-analytic review in the store environment. Journal of Business Research (66), 1420-1426.
Walsh, Gianfranco, Edward Shiu, Louise M. Hassan, Nina Michaelidou, and Sharon E. Beatty (2011) Emotions, store-environmental cues, store-choice criteria and marketing outcomes,
Journal of Business Research 64, 737-744.
Wu, J., Ju, H.W., Kim, J., Damminga, C., Kim, H-Y. & Johnson, K.K.P. (2013) Fashion product display. An experiment with Mockshop investigating colour, visual texture, and style
