The difficult choice what to choose

(1)

A study about the effect of assortment variety on purchase intention in Dutch supermarkets.

(2)

The difficult choice what to choose

A study about the effect of assortment variety on purchase intention in Dutch supermarkets.

Student: Tim Ruiter

Student number: S2025310

Faculty: Economics and Business

Master: MSc. Business Administration

Profile: Marketing Management

Qualification: Master Thesis

Completion date: 15-07-12

Adress: Olympiaweg 48A

1693 EL Wervershoof

Phonenumber: 06-13579072

Emailadress: timruiter88@gmail.com

Supervisor: Erjen van Nierop

(3)

Management Summary

A common assumption in marketing is that more varied assortments are beneficial for consumers. This assumption is consistent with the prediction by classic economic theories that larger, more varied, assortments should always be beneficial for consumers because they provide for a potentially better match between consumers' own preferences and the product offering. Contrary to the common wisdom that more choice is always better, other research suggests that limited choice is more preferable in some scenarios. Selections made from large varied assortments can lead to weaker preferences (Chernev, 2003a). The findings of Iyengar and Lepper (2000) are in line with other studies that suggests that increasing the amount of options in an assortment may have adverse effects on the strength of consumer preferences because it can confuse consumers, increasing the likelihood they will delay their choice or not choose at all (Kida et al, 2010). Furthermore, in a study of Iyengar and Lepper it was shown that consumers who faced a limited assortment had higher purchase intentions then other consumers who were confronted with an extensive varied assortment.

This study tries to analyze the relationship between assortment variety and purchase intention in the context of Dutch supermarkets. Research was conducted in four product category scenario's in Dutch supermarkets. Each product category symbolizes one category role described by Dhar Hoch and Kumer (2001).The four roles, staple products, niches, variety enhancers and fill-ins are based on the percentage of households buying and the frequency of purchase. Each product category: chips (staple product), baby nutrition (niche), tea (variety enhancer) and dried herbs and spices (fill-in) was investigated in a limited and an extensive varied assortment context.

Consumers and retailers recognize the value of variety in a store's assortment. Hoch et al, (1999) stated three reasons why consumers care about assortment variety: (1) Shoppers would rather go to a store and find and purchase exactly what they want, (2) Assortment variety offers the consumer option value and (3) consumers may care about variety because of an innate desire to consume different alternatives. Hoch et al. (1999) states that the relation between assortment variety and the products consumers buy is affected by their prior experiences, product knowledge, choice costs and their perception of assortment variety. According to Alba and Hutchinson (1987) the experience with and knowledge about a product resembles the expertise of a particular consumer. Prendergast et al. (2010) concluded that someone's personal decision involvement in a product is affecting their purchase intention for that product. So in conclusion, a persons’ perception of assortment variety, prior experiences and product knowledge (combined in product category

(4)

assortment variety, product category expertise, personal decision involvement and choice costs were used in an online questionnaire used for data collection.

The results indicate that consumers do not have higher purchase intentions when the assortment offers extensive variety in comparison with assortment that offers low.

Furthermore, perceived assortment variety positively affected purchase intentions directly in all product categories. Product category expertise showed a positive effect on purchase intentions directly in the product categories tea, dried herbs and spices and the generalized data of all four product categories in this sample group.

Perceived assortment variety showed a significant negative influence on the relation between assortment variety and purchase intention in the product categories chips and tea. This means that consumers who perceive high levels of assortment variety have lower intentions to purchase within that product category. Consumers might evaluate an extensive varied assortment as overwhelming or confusing, increasing the likelihood they will delay their choice or not choose at all (Kida et al, 2010).

Consumers with higher levels of personal decision involvement, who were confronted with an extensive varied assortment, showed significantly higher purchase intentions than those consumers who were confronted with a limited varied assortment in the product categories chips, tea, and the generalized data of all four product categories.

It is notable that choice costs did not have any significant influence on consumers purchase intentions or its relationship with assortment variety. Supposedly consumers do not experience high levels of choice costs when they shop for groceries at supermarkets and/or it is not influencing their purchase intention.

Other notable results indicate men are less likely to purchase tea. Furthermore, the results suggest that when someone's income increases, their purchase intention for dried herbs and spices decreases.

In summary, it can be concluded that extensive varied assortments do not improve purchase intentions. Further it appeared that personal decision involvement is positively effects the relation between assortment variety and purchase intention. Generalization upon these results has to be done carefully because of the differences across the analysed product categories and the

(5)

Preface

This thesis has been written on behalf of graduating for the master Business Administration, specialisation Marketing Management, at the University of Groningen. I am grateful that I had the opportunity to do my research and write this thesis to graduate.

I want to take this opportunity to thank everyone who contributed to this research and in particular the respondents who participated in the study. Without them, doing this research would have been impossible.

Of course I also want to thank my first supervisor from the University of Groningen, Erjen van Nierop. He guided me throughout this research, and with his guidance, critique and support this thesis was raised to a higher level. Furthermore, I would like to thank Jacob Wiebenga, my second supervisor. His feedback in the final stages of the process helped me to finish this report.

Lastly, I want to thank all my close friends and family for their support throughout the whole trajectory. Without them, writing this thesis would have been a whole lot harder.

(6)

Chapter 1 - Introduction

1.1 Background

For a long time it was assumed that larger, more varied assortments should always be beneficial for consumers because they provide opportunities for a potentially better match between consumers' own preferences and the product offering. It seems rather plausible that having more options to choose from is preferable to having fewer options from which to choose. However, the paradox of choice phenomenon suggests that this may not always be true (Kida et al, 2010). Recent studies showed that respondents exposed to limited, less varied offering were more satisfied with their choice and had a higher purchase intention than consumers exposed to extensive respondents.

This study focuses on optimizing assortment variety for different product categories in Dutch supermarkets. The definition of assortment variety in this study is the extent of different brands and brand items within an assortment.

Retailers strive to maximize profit while dealing with lower margins and fierce competition. Assortment variety has the potential to positively influence assortment performance and thereby affecting a retailers profit. All products available in supermarkets belong to certain product

categories. For example "Soft drinks" or "Chips". Besides a certain category, all available products can be allocated in one of four different roles based on the percentage of households buying and the frequency of purchase. (Dhar, Hoch and Kumar, 2001). The category roles are displayed in table 1.

Table 1

Category roles by Dhar, Hoch and Kumar

Percentage of households buying

Frequency of Purchase - High penetration - Low penetration

- High frequency Staple products (like coffee) Niches (like yoghurt)

- Low frequency Variety enhancers (like pickles) Fill-ins (like pancake mix)

The variety of an assortment could have a positive effect on assortment performance. Overall assortment performance can be monitored by category development index (CDI). CDI compares category development of a retailer with the category development of competitors Dhar et al. (2001) showed that enlarging assortment variety had a positive effect on unit and dollar CDI for variety enhancers, niches and fill ins, because larger, more varied assortment can meet more heterogeneous needs. Staple products seem to have no effect on unit and dollar CDI. This can be explained because staples already have large, varied assortments ending up with a ceiling effect. Assortment variety has the power to improve the product category performance (Dhar, Hoch and Kumar, 2001).

(8)

In particular, the factors that influence the relationship between assortment variety and purchase intention have not been studied before.

This research study tried to indentify and analyze the relation between assortment variety and purchase intention and to help retailers optimizing their assortment variety.

1.2 Problem statement and research questions.

The main research question that was investigated in this research is:

"What is the effect of assortment variety on purchase intention?"

The main research question is coupled with the following research questions: 1) What is already known about assortment variety, purchase intention and its relation? 2) Which factors influence the relation between assortment variety and purchase intention? 3) Which managerial implications can be derived to improve the product category performance via assortment variety?

1.3 Academic and practical relevance

The advantages and disadvantages of limited or extensive variety in an assortment has been discussed in many research studies (Lancaster, 1990; Ratner et al., 1999; Broniarczyk et al, 1998; Oppewal et al., 2005; Kida et al., 2010; Iyengar and Lepper, 2000; Iyengar, Jiang, and Huberman, 2004). However, there are hardly any detailed studies about the relationship between assortment variety and purchase intention. In particular, which constructs are moderating the relationship between assortment variety and purchase intention has not been studied before. Studying the relationship between assortment variety and purchase intention is very relevant in explaining a part of why consumers buy or do not buy products in certain occasions.

(9)

and to come up with more specific implications for retailers. Furthermore, it might contribute to the solution of the paradox of choice phenomenon.

The second contribution of this research is that the generalizability of the relationship between the extent of variety in an assortment and purchase intention across four types of product categories (chips, baby nutrition, tea and herbs & spices) is investigated. Each product category has a different category role, based on the percentage of households buying and the frequency of

purchase (Dhar, Hoch and Kumar, 2001). The potential moderating impact of perceived variety, product category expertise, purchase decision involvement and choice costs on the effect of

assortment variety on purchase intention is revealed. In doing so, recent research that addressed the relationship between assortment variety and purchase intention is extended.

This research will help to gain insights in the effects of the potential moderators on the relationship between assortment variety and purchase intention. In this way, this research hopes to add new theory and knowledge to the existing literature. Furthermore, the relationship between the extent of assortment variety and purchase intention in the context of product categories in Dutch supermarkets has not been investigated before.

The results of this research study give insights in how retailers could optimize their assortment variety to stimulate the purchase intention. By optimizing assortment variety, the

product category performance and profit will be positively affected. Assortments that offers too little variety could result in potential sales loss, offering too extensive varied assortments could result in choice overload and confusion, increasing the likelihood that consumers will delay their choice or not choose to purchase at all (Iyengar and Lepper, 2000) resulting in potential sales losses. Therefore, recommendations for retailers to improve their product category performance and profit by adjusting their assortment variety are very relevant.

1.4 Structure of the thesis

The remainder of this paper is structured as following. First, relevant theoretical concepts will be outlined to give a more clear insight on the research topic. Hypotheses derived from the literature review will be proposed at the end of the theoretical concept. The theoretical framework chapter will answer research question 1. In the methodology section the outline of the data collection and

(10)

Chapter 2 - Theoretical Framework

2.1 Purchase intention

Purchase intention can be defined as someone's intention to purchase a product. Purchase intention is often used as a tool to forecast sales. Purchase intentions have been widely used in the literature as a predictor of subsequent purchase (Grewal et al., 1998). Academic research showed considerable attention to the study of purchase intentions and the relationship between purchase intentions and actual behaviour (Chintagunta and Lee, 2012). The main reason for using purchase intentions measures is that there is some useful information contained in these measures that reflect consumers’ likelihood of purchasing a certain product. Managers are not interested in the likelihood of purchasing a certain product in itself but they use it as a tool to predict behaviour and thereby actual sales. Between the likelihood of purchase and the actual purchase is a time interval. In what way is this time interval affecting the actual behaviour?

The effect of the temporal separation between intention measurement and purchase on the ability of intentions to predict behaviour is addressed in several studies (Morowitz et al., 2007; Chintagunta and Lee, 2012). It was found that the smaller the time interval between the purchase intentions and the purchase, the better intentions can predict actual behaviour. This is in line with prior research by Fishbein and Ajzen (1975). When the time interval between the intentions and the purchase increases, the purchase intentions are more likely to change over time. The change in purchase intentions could come from internal sources, for example by a change in needs, or external sources like advertising or word-of-mouth across consumers. Under such circumstances, it is

reasonable to assume that intentions measured just prior to purchase are likely to be most diagnostic of a consumer’s actual actions since they reflect a greater amount of information than intentions measured well in advance of purchase (Chintagunta and Lee, 2012). This means that the smaller the time interval between the purchase intention and the actual purchase, the better the purchase intention predicts consumers’ actual behaviour.

This study tries to reveal the effect of assortment variety on purchase intention. The area of

purchase intention is addressed in many studies.Dodds et al. (1991) showed the effects of price,

brand and store information on consumers' purchase intention. But there is limited to non research available about the relationship between assortment variety and purchase intentions. What is known in literature about assortment variety?

2.2 Assortment variety

(11)

consumer's rank the variety of an assortment third right after price and location when they give reasons for patronizing their favourite store (Arnold et al., 1983). A study about consumer trends in supermarkets in 2011 of the EFMI business school stated that 'quality and assortment' is the leading motive in selecting a supermarket. So the assortment becomes even more important. Perceptions of variety are an important determinant of attitudes and store choice (Hoch et al., 1999). Variety is important because there is a positive relation between assortment variety and the fulfilment of consumers needs (Hoch et al., 1999). Simply said, consumers are more likely to find what they were looking for when they visit a store with more varied assortments. This means that assortment variety is important for retailers because consumers value variety.

Literature in this area suggests that consumers benefit with greater variety in almost all scenarios. However, for the retailer, this isn't the case. They have the difficult job to keep the demand side in balance with the supply side. Improving variety by adding more products in the assortment increases distribution, sales and operational costs (Balderston, 1956).

In conclusion it can be said that for both consumer and retailer the variety in a store's assortment is important.

A common assumption in marketing is that more varied assortments are beneficial for consumers. This assumption is consistent with the prediction by classic economic theories that larger, more varied, assortments should always be beneficial for consumers because they provide for a potentially better match between consumers' own preferences and the product offering. However, recent research suggested that this might not always be the case (Chernev, 2003a).

Extant research has identified that more varied assortments have the potential to increase the strength of consumer preferences. The most important aspect, featured often in literature, is that more varied assortments offer an opportunity for a better match between an individual's preferences and the characteristics of the alternatives in the choice set1_{(Lancaster, 1990).}

Consumers prefer more varied assortments because they might lead to stronger preferences because they offer value and flexibility (Ratner et al., 1999). Furthermore, consumers might

(12)

So it seems obvious that having more options to choose from is preferable over having fewer options from which to choose. However, the paradox of choice phenomenon suggests that this may not always be true (Kida et al., 2010). More specifically, while more options may provide consumers with the opportunity for a better decision outcome, the presence of an extensive number of options may result in choice overload and prove to be de-motivating. Extensive choice could result in a state of “decision paralysis” in which consumers simply choose not to choose (Iyengar and Lepper, 2000).

Large, varied assortments could lead to weaker preferences caused by the increased demand of consumers' cognitive resources for the extra effort required to evaluate all options in the

consideration set. Furthermore, increasing the size of the choice set might confuse consumers, leading to weaker preferences and lower choice probability (Dhar, 1997; Greenleaf and Lehmann, 1995). The paradox of choice might occur when consumers experience a higher level of anticipated regret for those alternatives that are not selected. Some alternatives may outperform the selected option. Furthermore, consumers may feel more personally responsible for an imperfect choice because with many options available, they could only blame themselves for an imperfect choice. Chernev (2003b) states that consumers often make choices with a lack of expertise, which is more easily accomplished when confronted with limited offering than extensive offering.

The paradox effect has been found in psychological research in several studies in different contexts. It was found in the participation in an extra credit assignment by students of the Columbia University (Iyengar and Lepper, 2000) and in the participation decisions by employees in employer-sponsored retirements plans (Iyengar, Jiang, and Huberman, 2004). This study of Iyengar, Jiang, and Huberman (2004) showed that employee participation rates were higher for retirement plans that offered fewer fund options than for retirement plans that offered more fund options.

In 2000, Invengar and Lepper analyzed the specific effects of choice-set size on the behavior of those who had to choose. They conducted two studies in different decision contexts, a retail context and at a university. The results suggests that the availability of extensive options may result in consumers' disability of identifying the best option available.

In the first study, participants were exposed to store displays that featured 6 or 24 flavors of jam in a specialty store. The study showed that a significantly larger percentage of shoppers exposed to the limited set size (30%) made a purchase than participants who were exposed to the extensive set size (3%).

(13)

extensive choice condition completed the assignment. The willingness to participate in an extra-credit assignment was higher for those assigned to the limited choice condition.

It can be concluded that consumers prefer to choose out of limited options in one scenario and out of extensive options in another. The expectation is that consumers would have higher purchase intentions when they are confronted with a more varied assortment because more varied assortments offers the opportunity for a better match between an individual's preferences and the characteristics of the alternatives in the choice set (Lancaster, 1990).

H1: Consumers have higher purchase intentions when the assortment offers extensive variety in comparison with assortment that offers limited variety.

2.3 The consideration set

Now, the concept of the consideration set, or choice set, is explained because it helps understanding how consumers derive at a purchase decision. The concept of the consideration set was introduced by Howard and Sheth (1969) and proven valuable in models of consumer response (Hauser and Wernerfelt, 1990). The basic idea is that consumers go through a, at least, two-stage process when deciding to make a purchase. In the first stage, they narrow down the global set of alternatives to a smaller set, the consideration set, from which a choice is made in the second stage (Van Nierop et al., 2010). Consumers confronted with a large number of brands use a simple heuristics to screen the brands to a relevant set called the consideration set (Alba and Chattopadhyay 1985). Purchase decisions are then made from brands in this set (Hauser and Wernerfelt, 1990).

The consideration set concept is consistent with a number of theories and results in behavioral science (Hauser and Wernerfelt, 1990). Wright (1975) found that consumers try to simplify their decision environment, Miller (1956) reported limitations on human abilities to process and store information and Alba and Hutchinson (1987) report several phenomena related to a simplification of choice through consideration sets. However, the view that consumers simplify decisions with consideration sets does not necessarily mean that consumers are lazy or that they are not rational. Such behaviour can occur when consumers' balancing of consumption utility and evaluation cost. The theoretical construct and definition of a consideration set in this paper is those brands and products that consumers consider seriously when making a purchase decision.

2.4 Possible moderators

(14)

Consumers and retailers recognize the value of variety in a store's assortment. Hoch et al, (1999) stated three reasons why consumers care about assortment variety:

1) Shoppers would rather go to a store and find and purchase exactly what they want. More varied assortments increase the probability of finding a perfect match (Baumol and Ide, 1956). This means that consumer's perceptions of variety effect their store choice and purchase decisions. This effect decrease when a purchase is routine, based on prior experiences, and consumers have

knowledge about what product they want and which stores carry it, which is often the case in buying groceries at a supermarket retailer.

2) Assortment variety offers the consumer option value. Perceived variety will matter more when preferences are uncertain (Hoch et al., 1999). Consumers who have never purchased in a particular category may seek education about available alternatives and/or will prefer visiting stores which are known for their high variety assortments to reduce search costs. The difficulty of selecting a product, the time and the extent of cognitive effort needed to make a choice, the choice costs, creates a situation in which consumers might prefer fewer options to choose from.

3) Finally, consumers may care about variety because of an innate desire to consume different alternatives. This perception of, and need for, assortment variety is driven by satiation and the need for stimulation achieved through exposure to novel stimuli.

Hoch et al. (1999) states that the relation between assortment variety and the products consumers buy is affected by their prior experiences, product knowledge, choice costs and their perception of assortment variety. According to Alba and Hutchinson (1987) the experience with, and knowledge about a product resembles the expertise of a particular consumer.

Prendergast et al. (2010) concluded that someone's personal decision involvement in a product is affecting their purchase intention for that product. Consumers’ purchase intention for a product with a low level of personal involvement should be lower than purchase intentions for a product with a high level of personal involvement. In other words, higher levels of personal decision involvement may imply higher purchase intentions. It seems obvious that consumers with high levels of personal decision involvement prefer extensive varied assortment because they contain more alternatives to choose from to derive at their favourite option.

In some scenario's consumers prefer extensive varied assortments because of their

(15)

In conclusion, could a persons’ perception of assortment variety, prior experiences and product knowledge (combined in product category expertise), their level of personal decision involvement and their level of choice costs help to explain the relationship between assortment variety and purchase intention? Before this question was answered, each of the potential moderators will be described in more detail.

2.4.1 Perceived assortment variety

According to Hoch et al. (1999) consumers may care about variety because of an innate desire to consume different alternatives. This variety seeking motive is driven by satiation and the need for stimulation achieved through exposure to novel stimuli. This means that consumer perceptions of the variety in an assortment are essential. A better understanding of the effect of perceived assortment variety on the relationship between assortment variety and purchase intention could help in solving the choice paradox. In some scenario's, consumer perceive a certain amount of assortment variety resulting in a preference for extensive varied assortments. In other situations, the perceived assortment variety may results in a preference for limited varied assortments. In studying the effect of perceived assortment variety, more knowledge is gained that could help to solve the choice paradox.

Previous research has shown that when a product category is given more physical space, so more facings, in a retail store, consumers perceive more variety than if it is assigned to a smaller space (Broniarczyk, Hoyer, and McAlister 1998). Van Herpen and Pieters (2002) stated that doubling the size of an assortment of replicated items increases the perceived assortment variety by as much as 42%. All things being equal, an increase in actual variety will increase perceived variety (Kahn and Wansink, 2004). Furthermore, an assortment's organization can influence perceived assortment variety (Hoch et al., 1999). For an extensive varied assortment, disorganization can make it more difficult for consumers to perceive and enjoy the full extent of the variety. On the other hand, when an assortment offers limited variety the organization of the assortment makes it relatively obvious that there are not many alternatives available. For limited varied assortments, disorganization can obscure the lack of alternatives and increase the perception of variety (Kahn and Wansink, 2004). It is the perceived variety of the assortment that influences consumption utility and ultimately

contributes to consumption quantity (Broniarczyk, 2004).The extent of perceived assortment variety and the satisfaction derived from having access to that variety could influence purchase intentions (Kohn and Wansink, 2004).

(16)

2.4.2 Product category expertise

The expertise of consumers with a product category consists of the experience with and knowledge about a product category (Alba and Hutchinson, 1987). Consumers’ experiences with a product category can vary from indirect to direct experience, depending on their level of interaction with products in a product category (Mooy and Robben, 2002). For example, product displays, being exposed to advertisement, or the stories of a vendor or other consumers can be seen as indirect product experiences because consumers did not fully interact with the product. They did not use the product itself. Product trials and usage give the user direct product experience. Prior research has compared the informational value of indirect and direct product experiences. Direct product experiences seem to have more value because they provide consumers with more credible

information than indirect experiences (Hamilton et al., 2007). Many consumers buy the same brand items over and over again. Habitual purchases are based on prior experiences and therefore part of product category expertise.

Besides experiences, product category knowledge is part of consumers’ product category expertise. A person's product category knowledge defines its capacity for decision making in that product category. When the product category knowledge increases, so does the capacity for decision making (Alba and Hutchinson 1987). Product category knowledge enables consumers to process new information more efficiently (Johnson and Russo 1984). When products are complex, consumers are faced with the task of identifying relevant attributes (Hoch and Deighton 1989). This seems easier for consumers with a relative high level of product category expertise. Experts are better in separating relevant from irrelevant attributes than novices (Larkin et al. 1980). They reduce the cognitive effort required for decision making (Bettman 1990). In addition, expert consumers have an easier time determining the attributes that resulted in need fulfilment during previous consumption experiences and can identify reasons for choosing one option over another more easily (Heitmann et al., 2007).

(17)

H3: For consumers with more product category expertise, the effect of assortment variety on purchase intention is stronger (more positive) than for consumers with less product category expertise.

2.4.3 Personal decision involvement

Someone's level of personal involvement in a decision about a product is affecting their purchase intention for that product. Someone's purchase intention is lower when their level of personal decision involvement is low. When their level of personal decision involvement is higher, their purchase intention should be higher. In other words, higher levels of personal decision

involvement may imply higher degrees of purchase intentions (Prendergast et al., 2010). The level of personal decision involvement indicates the willingness to cognitively process all available

information. Often, consumers with high levels of personal decision involvement for a product are more risk averse and rely more on factual and important information. Consumers with low levels of personal involvement on a particular product are less risk averse and more accessible for secondary cues like celebrity endorsement. Prendergast et al (2010) stated that this logic is supported by Petty& Cacioppo’s (1981) Elaboration Likelihood Model (ELM). During their daily life, consumers are

confronted with many persuasive messages each day. It is impossible to process all these messages. The ELM indicates that consumers can process the information from these messages in two different ways. They actively consider the information in a message, this way is called the central route, which requires a great amount of cognitive effort. Or they only attend to positive or negative cues in stead of argument strength in processing a message. This way is called the peripheral route. The ELM suggests that the pathway used in processing a message is moderated by the level of personal involvement. The level of personal involvement is determining the pathway used in processing the incoming message and the response to the message.

When a consumer has high levels of personal involvement with a product, a message is processed by the central route. The consumer will actively process the message with high levels of cognitive effort. Both product attributes and cues are processed to arrive at a purchase decision. On the other hand, when a consumer has low levels of personal involvement in a product, the consumer will be less motivated to actively process the message with high levels of cognitive effort. Product attributes seems less likely to be important in arriving at a purchase decision. Instead, a consumer with low levels of personal involvement in a decision for a product is more likely to base their purchase intention on peripheral cues such as the package (Prendergast et al., 2010).

(18)

involvement is moderating the response to the message, the response in the ELM is the purchase decision. The assumption is that consumers with high levels of personal decision involvement prefer more varied assortments because they are willing to invest cognitive effort in selecting the best offer available based upon product information and simple cues like the package. Consumers with low levels of personal decision involvement would base their choice upon simple cues like the package and in-store stimuli.

A part of the choice paradox phenomenon might be revealed by analyzing the effect of purchase decision involvement on the relation between assortment variety and purchase intention. The assumption is that consumers with high levels of personal decision involvement prefer more varied assortments because they are willing to invest cognitive effort in selecting the best offer available, which assumes that they prefer more varied assortments.

Lancaster (1990) stated that more varied assortments offer the opportunity for a better match between an individual's preferences and the available options in the assortment. A better match would probably lead to higher purchase intentions. Therefore the expectation is that for consumers with a high level of personal decision involvement, the effect of assortment variety on purchase intention is stronger, more positive, than for consumers with a low level of personal decision involvement.

H4: For consumers with a high level of purchase decision involvement, the effect of assortment variety on purchase intention is stronger (more positive) than for consumers with a low level of purchase decision involvement.

2.4.4 Choice costs

The choice costs consist of the difficulty of selecting a product and the time and the extent of cognitive effort needed to make a choice.

Consumers use different types of thought processes when they choose products. The

thought process in use depends on the product and characteristics of the choice set (Bettman, 1988). According to Lussier and Olshavsky (1979), in a choice set with a large number of products, a

consumer is likely to use a choice process that eliminates most of the products. The variations in choice processes could be explained by the fact that the decision maker is influenced by goal of minimizing cognitive effort or the mental effort of thinking (Bettman, 1988; Cooper-Martin, 1989; Bettman, Johnsen, and Payne, 1990).

(19)

products on a large set of attributes (Cooper-Martin, 1994). Humans are limited in their processing capabilities; that is, people try to conserve their limited resources and save effort (Heitmann et al, (2007). Decision making requires more effort when the number of options is large (Payne, Bettman, and Johnson 1992).

More choice makes it more difficult to make a decision, and such difficulties may lead to confusion, regret, disappointment and dissatisfaction (Schwartz 2004). The paradox of choice can occur because consumers could experience a higher level of anticipated regret for those alternatives that are not selected. Some alternatives may outperform the chosen option or consumers may feel more personally responsible for an imperfect choice.

Consumers' preference for limited variety in an assortment instead of extensive variety in certain contexts might be influenced by the level of effort needed to evaluate alternatives. Chernev (2003b) states that consumers often make choices with a lack of expertise, which is more easily accomplished when confronted with limited offering than extensive offering. Consumer behavior research discussing the adverse impact of assortment on choice focuses on the extra effort needed to evaluate the additional alternatives in a larger, more varied, assortment.

The fundamental premise on which the information-load paradigm is based is that

consumers have finite limits to absorb and process information during any given unit of time. Thus, if consumers are provided with "too much" information at a given time, such that it exceeds their processing limits, overload occurs leading to poorer decision making and dysfunctional performance (Malhotra, 1982).

Many companies compete by offering large, varied assortments. However, more varied assortment strategies can backfire if the complexity causes information overload such that a

customer feels overwhelmed and dissatisfied, or chooses not to make a choice at all (Jacoby, Speller, and Beming, 1974). The information overload may cause even bigger problems when companies try to implement customization, offering each client the possibility to choose exactly what they want. For example, www.tulpfietsen.nl offers customers the opportunity to select bicycle parts in 18 different part categories, which could end up in over a million different 'customized' bicycles.

Greenleaf and Lehmann (1995) found that the difficulty of selecting a single alternative was one of the most important causes for delaying a number of purchase decisions. So, the effect of offering greater choice needs to be in line with potential deferral due to increased customer confusion (Dhar, 1997).

(20)

If this expectation hold, it might help in solving the choice paradox phenomenon. Someone's level of choice costs could affect their preference for extensive or limited assortments. High levels of choice costs would indicate a preference for limited varied assortments.

In general, "too much" information to process could lead to an overload of information which may confuse consumers, delay them to choose, or make them decide not choose at all. It was

expected that consumers experience higher choice costs in selecting one item from an extensive varied assortment. Therefore, it was expected that the effect of assortment variety on purchase intention is weaker when consumers experience higher levels of choice effort.

H5: For consumers who experience high levels of choice costs, the effect of assortment variety on purchase intention is weaker (negative) than for consumers who experience low levels of choice costs.

2.5 Conceptual model

Could a person’s product category expertise, perceived assortment variety, choice effort, and personal involvement with a product or product category help to explain the relationship between assortment variety and consumers purchase intention? To set up a framework to answer this question, the conceptual model was developed. The conceptual model is displayed in model 1.

(21)

Chapter 3 - Methodology

3.1 Research method

Research was conducted in four product categories in Dutch supermarkets. Each product category was derived after discussion and symbolizes one category role described by Dhar Hoch and Kumer (2001).The four roles (Staple products, Niches, Variety enhancers and Fill-ins) are based on the percentage of households buying and the frequency of purchase. Each product category was investigated in a limited and an extensive varied assortment context resulting in eight conditions. One major brand was selected for each product category to avoid preferences between different brands. All selected brands are well known, market leader in their category and often purchased by consumers purchasing in that product category. For an overview of the category roles, product categories and the selected brands see table 2.

Table 2

Overview of category roles, product categories and brands Percentage of households buying

Frequency of Purchase - High penetration - Low penetration

- High frequency Staple products (Chips)  Lays Niches (Baby food)  Olvarit

- Low frequency Variety enhancers (Tea)  Pickwick Fill-ins (Dried herbs and spices) 

Verstegen

Respondents participated in an onlinequestionnaire with a 2 x 4 between subjects design.

There were two types of assortment variety, nm. limited and extensive. Furthermore four different product categories were used, nm. chips, baby food, tea and dried herbs and spices. The extant of variety within an assortment was manipulated, resulting in a limited and extensive varied assortment for all four product categories. Respondents were randomly assigned to one of the eight conditions. The questionnaire was the same for all eight conditions. In each condition respondents were confronted with a limited or extensive varied assortment and had to select one product item that favors them the most. When they selected their favorite product item they were asked to complete the questionnaire. Respondents were asked to complete the questionnaire after they selected their favorite product item from a given assortment to prevent them from being biased. Showing the questionnaire before selecting an option from a given assortment might bias the results.

(22)

For an overview of the assortment visualizations used in the questionnaire see appendix 1. For an overview of screenshot of the questionnaire in Dutch see appendix 2.

It was attempted to gather as many participants as possible. Each of the eight conditions should at least contain thirty respondents. Thirty respondents in each of the eight conditions results

in a minimum goal of 240 respondents. The actual number of respondents was321. In total, 248

respondents fully completed the online questionnaire which reflects a percentage of 77%.So, the

goal of thirty respondents per condition wasachieved.

3.2 Dependent variable

The dependent variable purchase intention is measured with a purchase intention scale. The scale measures the likelihood that a consumer will buy a certain product, also referred to as the

willingness to buy by Dodds, Monroe, and Grewal (1991). Purchase intention was measured by five

items: (1) "The likelihood of purchasing this product is", (2) "The probability that I would consider

buying the product is", (3) "My willingness to buy the product is" (4) "If I were going to buy ..., the probability of buying this flavour is", (5) "I would purchase this ... flavour". All these items are

measured on a 7-point Likert scale. The first four items ranged from 1 very low to 7 very high. Item five ranged from 1 strongly disagree to 7 strongly agree.

3.3 Independent variables

In the hypothesized model four independent variables were distinguished that could moderate the relationship of assortment variety on purchase intention: perceived assortment variety, product category expertise, purchase decision involvement and choice costs.

Perceived assortment variety was measured with the perceived assortment variety scale developed by Kahn and Wansink (2004) and intended to measure the degree of variety a consumer perceives there to be in an assortment of some product. Four items were used: (1) "This assortment

gives me a lot of variety", (2) "This assortment gives me at least one option I like", (3)"This assortment offers enough variety" and (4) "How much variety do you think there is in this

assortment?". These four items were measured on a 9-point Likert scale. The first three items ranged

from 1 strongly disagree to 9 strongly agree. Item four ranged from 1 low variety to 9 high variety. Product category expertise was measured to assess a person's expressed familiarity and experience with a certain category of products (Thompson et al., 2005). Five items were used on a 9-point Likert scale: (1) "How familiar are you with ..." ranging from 1 not familiar at all to 9 very familiar, (2) "How clear an idea do you have about which characteristics are important in providing

you maximum usage satisfaction?" ranging from 1 not very clear to 9 very clear, (3) "I know a lot about ..." ranging from 1 strongly disagree to 9 strongly agree, (4) "How would you rate your

(23)

one of the most knowledgeable people and (5) "How frequently do you purchase ...?" ranging from 1 never to 9 all the time.

Personal decision involvement was measured by the purchase decision involvement scale provided by Mittal (1989). Purchase decision involvement, the interest a person takes in a product at point of purchase, is very relevant for products sold in supermarkets. Four items were used on a 7-point Likert scale: (1) " In selecting from many brands and flavours of ... available in the market,

would you say that:" ranging from 1 would not care at all as to which one I buy to 7 I would care a

great deal as to which one I buy, (2) "Do you think that various brands and flavours of ... available in

the market are all very alike or all very different?" ranging from they are all alike to 7 they are all

different, (3) "How important would it be to you to make a right choice of ...?" ranging from 1 not at all important to 7 extremely important, (4) "In making your selection of ..., how concerned would you

be about the outcome of your choice?" ranging from 1 not at all concerned to 7 very much

concerned.

The choice costs were measured with the choice costs scale developed by Heitmann, Lehmann, and Herrmann (2007) and intended to measure the choice difficulty and level of time and effort expended during a purchase decision. Five items were used: (1) "How much time/effort did it

take to evaluate and compare the alternatives in order to feel comfortable making a choice?", (2) "I could not afford the time to fully evaluate relevant purchase options", (3)"It was tough to compare the different products being offered", (4) "It was difficult for me to make this choice" and (5) "I concentrated a lot while making this choice". These five items were measured on a 9-point Likert

scale. The first item ranged from 1 a lot to 9 very little. The remaining four Items ranged from 1 strongly agree to 9 strongly disagree.

An overview and definition of the variables including the measurement instruments is given in table 3.

Table 3

Overview and definition of variables

Variable Concept Measuring instrument

Purchase Intention The willingness to buy a

product. Five items, 7-point Likert scale (Chintagunta and Lee, 2012)

Perceived assortment variety The degree of variety a

consumer perceives there to be in an assortment of some product.

Four items, 9-point Likert scale (Kahn and Wansink, 2004)

Product category expertise A person's expressed familiarity

and experience with a certain category of products.

Five items, 9-point Likert scale (Thompson et al.,2005)

Personal decision involvement The interest a person takes in a

product at point of purchase. Four items, 7-point Likert scale (Mittal, 1989)

Choice costs The choice difficulty and level of

(24)

(25)

Chapter 4 - Empirical results

The aim of this study was to examine if the extent of variety in an assortment affected consumers' intention to purchase one of the displayed products. Literature assumed that the level of perceived assortment variety, product category expertise, purchase decision involvement and choice costs might moderate the relationship between assortment variety and purchase intention. Of course, these assumptions needed to be tested. First, the descriptive analysis will be discussed. The descriptive analysis will address the socio-demographic characteristics of the respondents who participated in this study. Then the results of the Cronbach's Alpha reliability analyses will be illustrated. Then the pooling test will be addressed. Furthermore, the results of the multiple regression analyses are explained.

4.1 Descriptive analysis

In total, 248 respondents completed the online questionnaire. The socio- demographic characteristics of this sample group are presented in table 4. In order to determine if the sample group was representative for the Dutch population, the results will be compared with the regular Dutch shopper population according to Centraal Bureau Levensmiddelenhandel (CBL).

Table 4

Overview of soc io- demographic characteristics

Demographic Variable (sample size) Sample group (248)

Regular Dutch shoppers CBL 2011 (1786) Sex

Female (%) 56.7 71

Male (%) 43.3 29

Age (in years)

16 – 25 (%) 20.8 26 – 40 (%) 21.7 41 – 55 (%) 27.5 56 – 70 (%) 25.0 Above 70 (%) 5.0 34 or below (%) 23 35 till 54 (%) 42 55 or older (%) 35

Education (based on Dutch system)

Lower (%) 15.8 24.2

Middle (%) 36.3 37.5

Higher (%) 47.9 38

Did not answer (%) 0.3

Total income per year (average is € 33.000) Not available

Below average (%) 13.3

On average (%) 34.6

1.5 times the average (%) 22.9

2 times the average (%) 10.4

More than 2 times the average (%) 7.9

(26)

Females were slightly more represented in the sample group than men (56.7% vs. 43.3%) but they were less represented than in the regular Dutch shopper population were females represented 71% of the shoppers. The age of the respondents was nicely distributed, only respondents above 70 were less represented. This might be caused by the lack of skills to participate in an online

questionnaire. These results are about the same as compared to the regular Dutch shoppers. Approximately, 15.8% of the sample group completed an lower education. About 36.3% enjoyed a middle level of education. Most people in the sample group enjoyed a high level of education (47.9%). In comparison with the regular Dutch shoppers, the sample group is highly educated. Furthermore, 13.3% had a total income per year below average, 34.6% on average and 40.9% above the average. 10.8% did not want the answer this question. There is no data about the income of the regular Dutch shopper group. Overall, the sample group contained slightly more females, had a nice distribution of age, had a relative high level of education and ditto income (40.9% above the

average).

The sample group is tolerable representative for the Dutch population but it could have been better. An overview of the mean scores other independent variables is shown in table 5.

The product category baby nutrition results showed the lowest purchase intentions of all product categories. The purchase intention scores for chips, tea and dried herbs and spices are more or less the same. Furthermore, the product category expertise of the product category baby nutrition is far lower than the other product categories. This is probably the case because respondents did not have a lot of experience with this product category. It is notable to see that the extensive chips assortment scored the highest product category expertise and perceived assortment variety scores and relative low on choice costs. An possible relation might be revealed in the multiple regression analyses. Respondents evaluated the highest choice costs in the product category baby nutrition.

Table 5

Overview of mean scores

Product category Purchase

intention Perceived assortment variety Product category expertise Personal decision involvement Choice costs General 4.99 6.14 4.92 5.18 4.06 General extensive 5.23 6.95 4.91 5.09 4.60 General limited 4.76 5.33 4.93 5.28 3.53 Chips 5.27 5.86 5.83 5.57 3.56 Chips extensive 5.66 7.01 5.91 5.54 3.90 Chips limited 4.88 4.71 5.74 5.60 3.22 Baby nutrition 4.08 6.28 2.78 4.75 4.84

Baby nutrition extensive 4.17 6.91 2.93 4.48 5.55

Baby nutrition limited 3.99 5.66 2.64 5.01 4.14

Tea 5.24 6.01 5.62 5.40 3.94

(27)

Tea limited 5.05 5.18 5.88 5.58 3.13

Dried herbs and spices 5.39 6.40 5.45 5.02 3.92

Dried herbs and spices extensive 5.67 7.03 5.43 5.12 4.20

Dried herbs and spices limited 5.12 5.77 5.47 4.91 3.63

4.2 Cronbach's alpha analyses

Although the items from the online questionnaire were derived from validated scales in literature, it is extremely important to investigate whether these items represent the defined construct (Malhotra, 2007). Cronbach’s Alpha is perfectly suitable in examining internal consistency between different items of a construct. The Cronbach's Alpha was used as a coefficient of reliability. When the Cronbach’s Alpha is high enough, the separate items together can represent the construct. The Cronbach’s Alphas are displayed in table 6.

Table 6

Overview of Cronbach's Alpha

Variable Cronbach's Alpha Number of items

Perceived Variety 0.843 5*

Expertise 0.902 4

Involvement 0.717 5*

Choice Costs 0.773 5*

Purchase Intention 0.941 5

* 1 item was deleted to improve the Cronbach Alpha The Cronbach’s Alpha for all of the constructs is found to exceed the .6 thresholds, which implies a sufficient level of scale reliability (Malhotra 2007). All Cronbach's Alpha scores have enough internal consistency to combine the scale items in five reliable variable scores. Item scores were summed and the total was divided by the number of used items to calculate the variable scores. For each variable score, higher scores indicated higher levels of that particular variable. From the construct scales of perceived assortment variety, purchase decision involvement and choice costs one item was deleted to improve internal consistency and reliability and thereby the Cronbach's Alpha score. The average of all items per construct was used to represent the construct in further analysis.

4.3 Pooling test

(28)

equation. In other words: combining the data is not allowed. Pooling was allowed in this study so the independent variables did not have different impacts in the different product categories. It was allowed to combine the data from all four product categories (.808 < 1.0099). For an overview of the calculation and an explanation of the pooling test see appendix 3.

4.4 Correlation analysis

Prior to multiple regression analysis it was checked if multicollinearity exists in the data. Therefore, the VIF scores among the independent variables were analysed. Table 7 shows that the VIF scores between the independent variables are high, which means that multicollinearity occurs (Sloot et al. 2005; Van Heerde et al. 2000). Moderation analyses is coupled with multicollinearity in almost all scenarios. It was assumed that centralizing the variables could lower the multicollinearity but recent research of Echambadi and Hess (2007) stated this was not the case. In this study, the multicollinearity can be explained by the amount of information in the dummy variable

assortment_extensive that is transferred to the interaction variables.

Table 7

Overview VIF scores independent variables

VIF

Perceived assortment variety 2,304

Assortment_extensive x perceived assortment variety 16,428

Product category expertise 2,278

Assortment_extensive x Product category expertise 10,761

Purchase decision involvement 2,313

Assortment_extensive x Purchase decision involvement 25,981

Choice costs 2,630

Assortment_extensive x Choice costs 11,265

Sex 1,123

Age 1,218

Income 1,212

Education 1,272

Assortment_extensive 40,520

4.5 Multiple regression analyses

Now the results of the multiple regression analyses will be discussed. To test the hypotheses , a multiple regression analysis with moderation was conducted. Moderation occurs when the relationship between two variables is influenced by a third variable, the moderator (Cohen et al., 2003). The moderators could affect the direction and/or strength of the relation between assortment variety and purchase intention.

(29)

The variety in an assortment in this study was limited or extensive making assortment variety a categorical variable. The extent of assortment variety was converted in dummy variable to make it suitable for a multiple regression analysis. The interaction variables were calculated by multiplying each of the constructs with the dummy variable assortment variety.

Moderation can be tested by conducting a multiple regression analysis with purchase intention as depended variable and the 13 independent variables: four possible moderators

individually, the interaction variables of all four possible moderators, the dummy variable assortment variety and the demographic variables sex, age, income and education level.

The significant moderating effect of a construct can be interpreted in this way: the main effect (b) of the independent variable indicates the effect of a limited varied assortment. The effect of an extensive varied assortment equals the sum of the main effect (b) plus the effect (b) of the interaction variable.

This result in the following multiple regression equation for this study:

Purchase intention = β0 + β1 perceived assortment variety + β2 product category expertise + β3 purchase decision involvement + β4 choice costs + β5 assortment variety + β6 (perceived assortment variety x assortment variety) + β7 (product category expertise x assortment variety) + β8 (purchase decision involvement x assortment variety) + β9 (choice costs x assortment variety) + β10 Gender +

β11 Age + β12 Income + β13 Education level +

Where perceived assortment variety, product category expertise, purchase decision involvement, choice costs, perceived assortment variety x assortment variety, product category expertise x assortment variety, purchase decision involvement x assortment variety, choice costs x assortment variety, Gender, Age, Income and Education level are the explanatory independent variables, β 0,1,..., β13 are the regression coefficients, and is the error associated with the regression.

First the general multiple regression analysis was conducted to gain insights derived from the combined data of all four categories. Then the data was analyzed with a multiple regression analysis per product category to explore the relation between assortment variety and purchase intention per product category and to reveal possible differences among different product groups.

4.5.1 General multiple regression analysis

(30)

adjusted R square score was used because it adjusts for the amount of independent variables. The Anova(b) was significant and showed that the independent variables are capable of predicting (a part) of the purchase intention.

Furthermore, perceived assortment variety, the interaction variable of perceived assortment variety, product category expertise and the interaction variable of personal decision involvement were the only variables that had significant influence on purchase intention.

Table 8

An overview of the results of the multiple regression analysis in general

Beta P-value

(Constant) 2.244 (p=.013)**

Perceived assortment variety .438 (p=.000)***

Assortment_extensive x perceived assortment variety -.191 (p=.043)**

Product category expertise .248 (p=.000)***

Assortment_extensive x Product category expertise -.120 (p=.232)

Purchase decision involvement -.051 (P=.626)

Assortment_extensive x Purchase decision involvement .631 (p=.000)***

Choice costs -.113 (p=.158)

Assortment_extensive x Choice costs .055 (p=.615)

Sex -.302 (p=.101)

Age .011 (p=.887

Income -.061 (p=.327)

Education .046 (p=.360)

Assortment_extensive -1.687 (p=.124)

p-value of the model .000

Adjusted R Square .400

* Significance at 10% ** Significance at 5% *** Significance at 1%

The results indicate that respondents who were confronted with extensive variety in an assortment did not had significant higher purchase intentions than those who were confronted with a limited varied assortment (p=.124). Therefore, hypothesis 1 is not supported. Someone's level of perceived variety in an assortment, their product category expertise, personal decision involvement or their choice costs might moderate the relationship between assortment variety and purchase intention.

(31)

The results indicate that perceived assortment variety effects someone's purchase intention (p=.000). But is this finding stronger for consumers who perceive higher levels of assortment variety? Although the moderation effect of perceived assortment variety on the relation of assortment variety on purchase intention is significant (p=.043) it has a negative effect (B -.191) and therefore weakens the positive effect of perceived assortment variety on purchase intention, as can be seen in figure 1. Hypothesis two is rejected. This finding is in line with recent studies (Chernev, 2003a; Kida et al, 2010; Iyengar and Lepper, 2000) and suggests that limited variety in an assortment can be

beneficial. Figure 1

Slopes moderating effect perceived assortment variety

0 1 2 3 4 5 6 1 7

Perceived assortment variety

Ef fe ct o n pu rc ha se in te nt io n Limited assortment Extensive assortment

How about product category expertise? Could different levels of product category expertise have effect on someone's purchase intention? Alba and Hutchinson (1987) suggested that the size and complexity of the consideration set, those brands and product that consumers consider seriously when making a purchase decision, is related to expertise within a category. Consumers with a

relative high level of product category expertise could deal with larger, more complex consideration sets. They could identify the best offering available relatively easy. Do consumers with a high level of product category expertise actually have higher purchase intentions? The significance of product category expertise (p=.000) indicates that the level of product category expertise effects the purchase intention. However, product category expertise does not have an moderating role in the relationship between assortment variety and purchase intention (p=.232). Hypothesis three is rejected.

(32)

cognitively process all available information. It was stated that more varied assortment provide a better match between preferences and the available options. This better match could lead to higher purchase intentions. On the other hand, more varied assortments could lead to choice overload resulting in lower purchase intentions.

Purchase decision involvement in itself does not have a significant effect on purchase intention in this study (p=.626). The moderating role of purchase decision involvement is significant (p=.000) with a high Beta of 0.631 and shown in figure 2. This indicates that level of involvement has a rather big impact on the relationship of assortment variety and purchase intention. Adding one unit of purchase decision involvement strengthens the effect of assortment variety on purchase intention by 0.631. Therefore, hypothesis four is accepted. This result confirms the assumption that higher levels of personal involvement stimulate the willingness to cognitively process all available information. The willingness to process all available information would result in a better match between an individual's preferences and the alternatives in the choice set in more varied assortment in comparison to limited varied assortments, resulting in higher purchase intentions.

Figure 2

Slopes moderating effect purchase decision involvement

0 1 2 3 4 5 1 7

Purchase decision involvement

Ef fe ct o n pu rc ha se in te nt io n Limited assortment Extensive assortment

(33)

needed to make a choice. Consumers experience different levels of choice costs in selecting one item from an assortment. It was expected that higher levels of choice costs would make consumers decide to delay their choice or not to choose at all. This would mean that higher levels of choice costs results in lower purchase intentions. The results indicate that this assumption is not correct. Choice costs do not have a significant effect on purchase intentions in this study (p=.158). The moderating role of choice costs in the relation of assortment variety on purchase intention is not significant either (p=.615). Perhaps consumers do not experience high level of choice costs but rather make decisions in supermarkets based on simple heuristics (Iyengar and Lepper, 2000). Furthermore, these results might have been biased because respondents did not had the opportunity to delay or cancel their choice because they were instructed to make a choice. There is no significant effect of choice costs on purchase intention and its relation with assortment variety. So, hypothesis five is rejected. 4.5.2 Multiple regression analysis per product category

Now, the results of the multiple regression analysis per product category are addressed. Each product group is analyzed individually to explore the relation between assortment variety and purchase intention per product category and to reveal possible differences among different product groups.

Chips

(34)

Figure 3

-4 -3 -2 -1 0 1 2 3 4 1 7

Purchase decision involvement

E ff ec t o n p u rc h as e in te n ti o n Limited Assortment Extensive Assortment Table 9

An overview of the results of the multiple regression analysis for product category chips Beta P-value

(Constant) -.201 (p=.938)

Perceived assortment variety .376 (p=.007)***

Assortment_extensive x perceived assortment variety -.115 (p=.580)

Product category expertise .219 (p=.331)

Assortment_extensive x Product category expertise -.146 (p=.662)

Purchase decision involvement -.462 (P=.128)

Assortment_extensive x Purchase decision involvement .712 (p=.032)***

Choice costs -.095 (p=.635)

Assortment_extensive x Choice costs .077 (p=.760)

Sex -.152 (p=.703)

Age -.123 (p=.501)

Income -.069 (p=.643)

Education .089 (p=.435)

Assortment_extensive 1.443 (p=.627)

p-value of the model .007

Adjusted R Square .268

(35)

Baby nutrition

The product category baby nutrition is a niche. Baby nutrition is bought by a few households, but the households that do buy baby nutrition buy on a regular basis as long as necessary. A

summary of the results of the multiple regression analysis for product category chips is shown in table 10. The analysis had an adjusted R square score of 0.202, which is the lowest of all four product categories. The Anova(b) was significant (p=.029). Similar to the product category chips, perceived assortment variety effects the purchase intention (p=.056). Purchase decision involvement

moderates the relationship between assortment variety and purchase intention (p=.052). Adding one unit of purchase decision involvement would result in .652 improvement in purchase intention. As figure 4 illustrates, the purchase decision involvement of consumers has a big impact on purchase intentions in the product category baby nutrition. This might be the case because most parents are very protective and cautious with their babies. High levels of purchase decision involvement indicate high levels of willingness to find what is best for the baby. When they find a product that meets their demands they would presumably want the buy that product. Hypotheses 1, 2, 3 and 5 are not supported in the product category baby nutrition. Only hypothesis 4 is supported, purchase decision involvement strengthens the relation between assortment variety and purchase intention in this product category.

Figure 4

0 1 2 3 4 5 6 7 1 7

Purchase decision involvement

The difficult choice what to choose