Online Second-Hand Shopping. Threat or Opportunity for Branded Products?

Table of Contents

List of Figures

List of Tables

List of Abbreviations

List of Symbols

List of Appendices

1 Introduction

2 Conceptual Background
2.1 Definition of Second-Hand Products
2.2 Characteristics of Second-Hand Products
2.3 Empirical Findings Related to Second-Hand Products
2.4 Consumer Brand Perceptions
2.4.1 Brand Image
2.4.2 Perceived Quality
2.4.3 Perceived Prestige
2.5 Product Availability
2.6 Theories on Scarcity Effects
2.7 Evaluation of Scarcity Theories regarding Applicability to Research Context

3. Development of Hypotheses
3.1 Main Effect
3.2 Moderating Effect of Product Type
3.3 Moderating Effect of Price
3.4 Interactive Moderating Effects of Product Type and Price
3.5 Control Variables
3.6 Descriptive Variables
3.7 Conceptual Framework

4 Empirical Study
4.1 Pre-Test
4.1.1 Stimuli Selection
4.1.2 Pre-Test Design and Results
4.2 Experimental Design
4.3 Procedure
4.4 Operationalization of Variables and Scenario Description
4.5 Data Cleaning and Sample Characteristics
4.6 Scale Assessment
4.7 Statistical Technique and Testing of Assumptions

5 Results
5.1 Manipulation and Realism Check
5.2 Testing of Hypotheses
5.3 Discussion of Results

6 Conclusion
6.1 Summary
6.2 Theoretical Contribution and Managerial Implications
6.3 Limitations and Further Research

Appendix

References

List of Figures

Figure 1: A Comparison of Cues for a New and a Second-Hand Product

Figure 2: Correlations of Selected Brand Perceptions in Consumer's Psyche

Figure 3: Types of Limited Availability

Figure 4: Model of the Scarce = Expensive Heuristic

Figure 5: Conceptual Framework

List of Tables

Table 1: Distinguishing Second-Hand Products from Other Types of Goods

Table 2: Reviewed Studies Related to Second-Hand Products and -Shopping

Table 3: Assessment of Selected Theories on Scarcity Effects

Table 4: Selected Products for the Pre-test

Table 5: Scale Items to Measure Hedonic and Utilitarian Features

Table 6: Hedonic and Utilitarian Goods Classification

Table 7: Operationalization of the Degree of Second-Hand Product Availability

Table 8: Overview of Used Scales for the Dependent Variables

Table 9: Sample Characteristics

Table 10: Content Validity and Reliability Statistics (N=562)

Table 11: Multivariate Analysis of Variance and Univariate Results

Table 12: Simple Contrast Results (K Matrix)

Table 13: Means and Standard Deviations

List of Abbreviations

illustration not visible in this excerpt

List of Appendices

Appendix A: Informal Analysis of Ebay.de for Each Pre-Tested Product

Appendix B: German Online Pre-test Questionnaire (Screenshots)

Appendix C: Pre-test Results

Appendix D: German Online Main Study Questionnaire (Screenshots)

Appendix E: Visualization of Product Type And Price Stimuli

Appendix F: Overview of Employed Scales (without Dependent Variables)

Appendix G: Data Cleaning (Main Study)

Appendix H: Example of Survey Promotion

Appendix I: Descriptive Statistics for Control and Descriptive Variables

Appendix J: T-Tests

Appendix K: Results of Factor Analysis

Appendix L: Evaluation of Formative Construct (Brand Image)

Appendix M: Reliability Statistics for Attitude towards Second-Hand Products

Appendix N: Testing of MANCOVA Assumptions

Appendix O: Results of Manipulation and Realim Check

Appendix P: Output of MANOVA

1 Introduction

Not long ago, browsing at flea markets and shopping at the corner second-hand store was considered as ‘curious’ niche activity (Guiot and Roux 2010; Sherry 1990a). However, the character of shopping for used goods has changed over the past years (Behrendt, Blättel-Mink, and Clausen 2011, p. 1). Due to recent techno- logical and societal developments, second-hand shopping has become a mass phenomenon (Guiot and Roux 2010; Hennig-Thurau et al. 2010). Although there is a lack of official statistics that measure the second-hand phenomena in terms of value and volume, various convergent estimations and recent surveys demonstrate that consumers are exhibiting a preference for second-hand products in ever- increasing numbers (Roux and Guiot 2011, p. 430; Sempora Consulting 2012)

The internet can be considered as major driver in this regard. Today, it seems that every product - however unique - can be found online anytime at an affordable price. This is the starting point of the long tail concept (Anderson 2004), which argues that when choice becomes apparently unlimited, demand moves from mass markets to millions of niches. The internet’s low search and transaction costs (Ghose, Smith, and Telang 2006) have also facilitated consumer-to-consumer (C2C) transactions. The biggest C2C platform eBay, where many second-hand products are traded, reports 124 million active users worldwide with 18 million users in Germany (eBay 2013a, eBay 2013b)

Another cause for the increasing popularity of second-hand shopping is the great- er orientation to sustainability among consumers (Brengman, Geuens, and Faseur 2002). Moreover, the negative social stigma of second-hand shopping, with its connotations of low income and economic necessity, is fading parallel to the rise of positive trends such as the popularity of vintage commodities (Cervellon, Car- ey, and Harms 2012). As demand for second-hand products has increased, so too has supply. About 45% of surveyed German consumers state that they resell goods (Sempora Consulting 2012). In addition, start-up companies are progres- sively investing in the so-called ‘pre-loved’ fashion industry by developing C2C second-hand platforms such as kleiderkreisel.de (Janke 2013). Such platforms of- ten concentrate on branded used products, particularly in the luxury goods catego- ry (Rose 2013).

These developments have fostered concern among practitioners that second-hand -fferings pose a threat to the perception of their brands (see here and in the fol- lowing, Sicking 2012). Brand manufacturers (for example adidas) fear that resell- ing on eBay could damage their brand’s image due to inconsistent price levels and product presentations. Consequently, some brand manufacturers try to prohibit online sellers from trading their products on eBay. Suitable to this observation, Hennig-Thurau et al. (2010) have suggested that future research should consider the impact of purchasing second-hand products on brand perceptions. These per- ceptions comprise all information related to the brand (Boltz and Leven 2004, p. 485). In this context, it seems that we also need to ask whether the information that branded second-hand products are actually available in itself alters brand per- ceptions.

Product type as a moderating variable has been used in prior research on product availability (Ku et al. 2013). This suggests that we should also examine whether the effect of the availability of second-hand products on brand perceptions is product-specific. Moreover, products are offered within a wide price range, so it would also be interesting to investigate whether high and low priced products are affected differently by the availability of second-hand products. In summary, the goal of our study is to investigate whether the degree of availability of second- hand products influences consumer brand perceptions while also considering the effects of product type and price.

Previous marketing studies related to second-hand products mostly deal with con- sumers’ motivations for buying used products (Guiot and Roux 2010; Roux and Korchia 2006). Micro- and macro-economic research discusses the extent to which the demand in second-hand markets may impact demand for the new prod- uct and if second-hand goods cannibalize sales of new goods (Anderson and Ginsburgh 1994; Ghose, Telang, and Krishnan 2005). Research in the context of second-hand shopping, products and markets largely neglects brand-related as- pects, although the literature highlights the importance of consumer brand percep- tions to marketers (for example Low and Lamb 2000). Studies on product availa- bility have predominantly been undertaken in the context of limited product avail- ability (scarcity) and resulted in several scarcity theories (Brehm 1966; Brock 1968). Despite its practical relevance and potential for a better understanding of consumers, no research exists that examines the relationship between the degree -f availability of second-hand products and consumer brand perceptions. Hence, this study contributes to the marketing literature as the first empirical research in- vestigating the impact of the degree of second-hand product availability on con- sumer brand perceptions while also considering the possible moderating effects of product type and price. The scope of the thesis lies within C2C-exchanges of orig- inally branded second-hand products which are traded through any act against payment. Additionally, we focus on (semi)-durable used goods, while leaving out cars, machinery and real estates. The results of the study could help brand manu- facturers to understand consumer perceptions and to clarify whether the degree of second-hand product availability actually influences brand perceptions. Our find- ings may also have managerial implications with regard to brands that focus on different product types or different price segments.

This thesis is structured as follows: First, we give an overview of the theoretical background related to second-hand products/-shopping, brand perceptions and product availability. These theoretical considerations as well as ideas from related research fields will then serve as a basis for the hypotheses development. We will also consider the influence of product-based moderators, i.e. product type as well as price, and present the conceptual framework of the study. This initial section is followed by a description of the empirical study whose results are subsequently reported and discussed. After summarizing our findings, the theoretical contribu- tion of the study and its managerial implications are outlined. We also discuss the study’s limitations and propose future research possibilities.

2 Conceptual Background

2.1 Definition of Second-Hand Products

Various authors have tried to formulate a concrete definition of second-hand goods in general or for specific product groups (for example Fuhrmann 1987, p. 6; Kornhardt 1984, p. 1; Meinig 1993, p. 96). Most of these attempts argue that sec- ond-hand goods can be defined with regard to the attribution of being ‘used’ or ‘previously owned’.¹ However, these simple definitions contribute little to clarify- ing the term as a whole, capturing its different facets and differentiating it from related product types. Due to the broadness of the concept of second-hand, it is necessary to distinguish second-hand goods from other and related types of prod- ucts such as brand new (Guiot and Roux 2010), fit for the scrap heap, antiquarian (Ohlwein 1990, p. 25), collectible (Guiot and Roux 2010) and recycled goods (Stroecker 1995, p. 7). One relatively new concept that is often linked to second- hand is vintage. It has been argued that the labels second-hand and vintage are close to become synonymous in consumers’ minds (Cervellon, Carey, and Harms 2012) whereas this study argues upon the very distinction between second-hand and vintage. The term vintage is primarily used in the fashion world, where it is defined as an authentic and rare piece that stands for the style of a particular era (Gerval 2008, p. 22) dating back to the period between the 1920’s and the 1980’s (Cervellon, Carey, and Harms 2012). Table 1 provides a ‘checklist’ for identifying second-hand products by applying three criteria to the depicted product types. As the original checklist by Ohlwein (1990, p. 32) only contains three product types other than second-hand, we have also included vintage and recycled products as well as collector’s items. It should be emphasized that the distinction between these seven types of products always bases on consumers’ subjective interpreta- tions.

Table 1: Distinguishing Second-Hand Products from Other Types of Goods.

illustration not visible in this excerpt

Source: Own Illustration Adapted from Ohlwein 1990, p. 32.

A defining characteristic of a second-hand good is the physical deterioration of the product, meaning that with every usage the product’s quality is reduced (Ohl- wein 1990, p. 29). By comparison, vintage goods are not necessarily deteriorated or used (Cervellon, Carey, and Harms 2012). Moreover, a second-hand product is still usable for its original purpose, which distinguishes it from a product that is fit for the scrap heap. Lastly, the third defining characteristic that distinguishes a second-hand good from, for example, an antiquarian or a vintage one is that it does not necessarily generate significant benefits from age-related shortages. Ac- cordingly, prices for vintage items can exceed those of luxury modern new pieces (Cervellon, Carey, and Harms 2012). This also applies to antiques and collectors’ items due to their rarity properties and the fact that these products are usually not available on the new goods market anymore. When referring to these products, Scitovsky (1994, p. 40) talks about “second-hand markets for the rich”. However, in this thesis we assume the co-existence of the branded second-hand product and the new product equivalent, which eliminates possible rarity properties of the sec- ond-hand good.

In summary, there is a need for a narrow definition to reduce the complexity asso- ciated with the term second-hand in order to gain an unambiguous understanding of its product characteristics. The working definition for the current thesis com- prises the following negative definition adapted from Ohlwein (1990, p. 25): A second-hand product is present as long as the product in question is not brand new, fit for the scrap heap, antiquarian, recycled, vintage or a collectible.

Throughout this thesis, we employ a broad product definition according to which a product is also regarded as a brand (Hogg, Cox, and Keeling 2000). Thus, when we use the term second-hand product, a branded product is assumed.

The next chapter discusses the characteristics of a second-hand product in more detail. This will enable us later to draw parallels to related research areas that fo- cus on the isolated effects of these product characteristics on brand perceptions.

2.2 Characteristics of Second-Hand Products

As soon as a new product’s technical, economic, and social lifetime is subjective- ly perceived as ending, the owner can discard it and sell it second-hand (Stroecker 1995, p. 17). The economic lifetime expires when technically improved models are available. The end of a social or psychological lifetime implies that a “con- sumer wants to have the newest of the newest (lifestyle)” (Stroecker 1995, p. 17). A group of people willing to purchase these disposed products create a second- hand market (Stroecker 1995, p. 13). Another important condition for a good to be qualified as second-hand is the product characteristic of (semi-)durability (Purohit 1992). Semi-durables have a technical life of between six months and three years, pure durables three years or more (Stroecker 1995, p. 19). Thus, almost every (semi-)durable can be offered second-hand (Müller 2009).

According to the cue utilization theory (Richardson, Dick, and Jain 1994), con- sumers rely on product cues when evaluating products. Those product cues can be divided into intrinsic and extrinsic ones (Olson and Jacoby 1972). The latter are product-related characteristics such as brand name, price, packaging and warran- ties that are not elements of the physical product (Richardson, Dick, and Jain 1994). Intrinsic cues comprise product-related attributes such as the design, per- formance, size and quality of ingredients (Bellizzi et al. 1981) that cannot be ma- nipulated without changing the physical product features. There are several prod- uct cues operating in new-product markets that are altered in second-hand mar- kets. Second-hand products are usually not packaged and the original packaging and/or the product tag does not exist anymore. Furthermore, warranties are large- ly missing for second-hand products (Guiot and Roux 2010) implying a higher risk when acquiring these goods (Stroecker 1995, p. 25). Moreover, the expected performance of a used product is lower in comparison with a new product (Stroecker 1995, p. 25). According to Srinivasan, Jain, and Sikand (2004), the product’s quality represents an intrinsic cue. Due to the natural usage-based quali- ty deterioration of the second-hand product, the objective quality of a second-hand product is lower compared to a new product (Stroecker 1995, p. 25). However, the physical deterioration varies from product to product as a result of the disparity in previous usage conditions (Ghose 2009). Thus, the second-hand product’s quality depends on product age as well as on the intensity and degree of use (Stroecker 1995, p. 25) and can be identified as a product-specific dimension. Thus, due to this quality heterogeneity (Brengman, Geuens, and Faseur 2002; Fox 1957), cus- tomers often remain in doubt regarding the quality of second-hand products. The most striking and obvious difference between a second-hand and a new product is the extrinsic cue of price (Stroecker 1995, p. 31). A used good is usually lower priced compared to a new product (Stroecker 1995, p. 22). However, due to the heterogeneous quality or physical condition of the goods, the same products can attract different prices (Stroecker 1995, p. 24) and bargaining is possible (Stroecker 1995, p. 25).

Although differences exist, second-hand products are sometimes viewed as (im- perfect) substitutes for new products (Stroecker 1995, p. 18). The brand name of a second-hand product is the same as for the original new product. Still, as Gabbott (1991) states, due to the uncertainty about the second-hand product’s quality and an associated higher functional risk, a brand name is a less reliable cue in the sec- ondary than in the primary market. As illustrated in Figure 1, the previous usage history of the products adds an extra dimension to both the extrinsic and intrinsic cues of a second-hand product (Gabbott 1991). This is in keeping with the fact that a second-hand product cannot maintain a first-owner status in the buyer’s life (Roux and Guiot 2008). Figure 1 illustrates the product cues elaborated above for a new good in contrast to a second-hand good.

Figure 1: A Comparison of Cues for a New and a Second-Hand Product.

illustration not visible in this excerpt

Source: Own Illustration.

2.3 Empirical Findings Related to Second-Hand Products

Researchers who deal with the topic of second-hand products and second-hand shopping repeatedly highlight the fact that it is a retailing phenomenon which has received little attention in the literature despite its long tradition (Guiot and Roux 2010; Roux and Korchia 2006). In order to verify these statements, to create an -verview of the current state of research and to identify findings relevant to our underlying research aim, we have included a literature review on empirical studies related to the second-hand context. The preceding chapter on second-hand product characteristics is mainly based on insights drawn from micro- and macro- economic literature. It is particularly second-hand markets that are discussed in this literature (for example Anderson and Ginsburgh 1994, Scitovsky 1994) and findings are derived from mathematical and stochastic models. However, in this thesis we focus on empirical studies carried out in marketing, psychology, and consumer research, since our research focus is on understanding consumers and their brand perceptions, an issue obviously more closely related to these research fields. We consider articles published in leading academic journals from the mar- keting, consumer and psychology fields that cover the time span from 1990 till 2013.² This literature restriction has been applied because the articles that were published before 1990 mainly focus on the economic advantages of second-hand shopping (Roux and Guiot 2008) and ignore the recreational benefits stressed from the 1990’s onwards. Additionally, we judge the findings before 1990 to be outdated for our research purpose. This is due to societal and technological trends of the present time such as underlying changes to the image of second-hand prod- ucts in combination with vintage commodities and the rise of the internet, which has become particularly important for second-hand shopping. Table 2 presents the reviewed studies’ key findings and the second-hand setting if specified.

Table 2: Reviewed Studies Related to Second-Hand Products and -Shopping.

illustration not visible in this excerpt

When analyzing the contents of Table 2, it becomes obvious that most of the stud- ies deal exclusively with offline settings such as car boot sales and flea markets while ignoring the online second-hand market. It also lacks studies of how con- sumers perceive second-hand products, especially in the branding context. The majority of studies focus instead on exploring motivations of second-hand shop- ping and channel specific characteristics. Apart from Roux and Korchia’s study (2006), the elicitation of attitudinal and perceptual aspects of second-hand prod- ucts is sparse. However, since it is these aspects that have the most relevance to the thesis’ research aim, only findings from Roux and Korchia (2006) are present- ed in more detail below.

The exploratory study by Roux and Korchia (2006) investigates the psychological aspects and the symbolic meanings associated with second-hand products. How- ever, their considered product category is second-hand clothing and therefore con- tains limited information. Roux and Korchia (2006) reveal that consumers have specific attitudes of refusal or acceptance towards second-hand clothes. Some people “perceive used clothing essentially as rubbish” (Roux and Korchia 2006, p. 33) and hold the view that clothes can only be worn by a single owner. In such cases, second-hand products are thus seen as highly contaminating due to the fact that a possible negative physical (concerns in hygiene) or symbolic transfer from the (unknown) previous user can occur (Argo, Dahl, and Morales 2006). These findings highlight the fact that contamination is a major factor influencing the re- jection of second-hand clothing, especially for those who have a high degree of attachment to their clothing products (see here and in the following, Roux and Korchia 2006). Thus, rejection and acceptance behaviors towards second-hand products depend on whether an individual considers a product as an element of someone else’s extended or their own self rather than as mere objects. When used goods are not viewed as an intimate part of the self or their extended self, they are appreciated for their intrinsic properties.

After having illuminated the second-hand phenomenon which constitutes the con- text of this thesis, the next section deals with the brand perceptions selected for our study.

2.4 Consumer Brand Perceptions

Since brand perceptions are the affected construct in our research aim, selected brand perceptions will be explained in the following, namely brand image, per- ceived quality and perceived prestige. Brand perceptions themselves are employed as a blanket term referring to the three selected constructs for measuring how con- sumers perceive brands. Our three brand perception constructs have been chosen because they are all cited as being perceptual components of brands in prior mar- keting literature and are all found to have different and reliable measures. The se- lected brand perceptions are not understood as independent entities. We rather suggest that to a certain extent they are correlated and overlap in the psyche of consumers. This is illustrated in Figure 2. However, the selected constructs do not encompass all aspects of brand perceptions due to the highly complex nature of consumer perceptions and can thus not be understood as an aggregated measure (Low and Lamb 2000). For this reason the following sections give an understand- ing and definition of the selected components of brand perceptions.

Figure 2: Correlations of Selected Brand Perceptions in Consumer's Psyche.

illustration not visible in this excerpt

Source: Own Illustration.

2.4.1 Brand Image

It is important to have a common understanding of the concept of brand image since researchers do not agree on a consistent definition and operationalization (Dobni and Zinkhan 1990; Low and Lamb 2000). However, since Keller’s (1993) brand image conceptualization is well-established in branding literature, this study adopts his definition. Brand image is thus defined as “perceptions about a brand as reflected by the brand associations held in consumer memory” (Keller 1993, p. 3). Specifically, brand image is regarded as the combined effect of brand associations and is a multi-dimensional and/or summary construct which comprises much brand information (Jacoby, Olsen, and Haddock 1971). Brand associations com- prise the meaning of the brand for consumers (see here and in the following, Kel- ler 1993). Their underlying dimensions are favorability, strength and uniqueness. Brand associations can be categorized into three types, namely attributes, benefits and attitudes. Attributes are the descriptive properties that typify a product and constitute the most objective level of associations. The personal values that con- sumers attach to the product are known as benefits or utilities. These can be fur- ther distinguished into functional, experiential and symbolic benefits. Lastly, brand attitude is linked to beliefs about functional and experiential benefits as well as to product-related attributes. Brand image is said to change continually and re- flects the latest brand perceptions (de Chernatony 1999). Thus, brand image can- not be described as static (Martinez and de Chernatony 2004) and is influenceable by many factors (Ueltschy 1998).

2.4.2 Perceived Quality

Perceived quality is one of the most commonly quoted consumer brand percep- tions in the marketing literature (Low and Lamb 2000; Yoo, Donthu, and Lee 2000). It is often defined as the consumer’s on judgments about the product’s su- periority (Zeithaml 1988) relative to alternative brands (Low and Lamb 2000; Netemeyer et al. 2004). In this thesis, we narrowly define perceived quality as the product’s superiority with regard to the product’s durability, reliability and per- formance dimensions. Durability comprises the length of time the product lasts as well as the length of time the product works (Brucks, Zeithaml, and Naylor 2000). Performance which also includes reliability relates to how well the product does what it is required to do (Brucks, Zeithaml, and Naylor 2000). The supplement ‘perceived’ of perceived quality makes clear that it is a rather high-level abstrac- tion and subjective component (Zeithaml 1988). Conversely, objective quality de- scribes the actual (technical) excellence or superiority of the product (Hjorth- Anderson 1984) which is verifiable or measurable according to certain established standards (Zeithaml 1988).

2.4.3 Perceived Prestige

There are several dimensions of a brand’s perceptual value that have been identi- fied in the literature (Sweeney and Soutar 2001). Our study concentrates on more social and hedonic aspects of the perceived brand value (for example Sweeney and Soutar 2001; Vazquez, Del Rio, and Iglesias 2002), namely the perceived brand prestige (Baek, Kim, and Yu 2010). The latter concept can be defined “as the relatively high status associated with a brand” (Baek, Kim, and Yu 2010, p. 663). The symbolic meaning which may be embedded in brands therefore repre- sents the prestigious value of a brand (Steenkamp, Batra, and Alden 2003).

Higher prices (Lichtenstein, Ridgway, and Netemeyer 1993) as well as the impact on reference groups (Bearden and Etzel 1982) often indicate the prestige of a brand (see here and in the following, Baek, Kim, and Yu 2010). Brand prestige, however, does not impact every consumer to the same extent. Publicly ‘insecure’ consumers in particular are concerned about how others perceive them and are rather prestige sensitive. Consequently, if a brand is prestigious and owned by a consumer it can enhance the consumer’s self-expression and social standing (O’Cass and Frost 2002).

While brand image and perceived quality are commonly used as consumer per- ceptions in the literature (Low and Lamb 2000), perceived brand prestige is a more ‘special’ brand perception. However, previous research emphasizes that it is an important virtue of a brand to possess a high (symbolic) perceived value (for example Sweeney and Soutar 2001), which also included the value of perceived prestige (Steenkamp, Batra, and Alden 2003). This is why this construct has been chosen for our research aim.

The next section deals with the assumed influencing factor on brand perceptions, i.e., product availability which will be embedded in the second-hand context.

2.5 Product Availability

The availability of a product can be defined as how much of a product exists rela- tive to a zero state (Brock 1968, Worchel, Lee, and Adewole 1975). By adopting this definition, our thesis follows prominent researchers in the field of product availability such as Brock (1968) and Brehm (1966), who commonly employ the zero state as the baseline when discussing product availability (Worchel, Lee, and Adewole 1975).

Evoking the thesis’ underlying research aim, we will examine the effects of the degree of second-hand product availability (stimuli) on consumer perceptions. We explicitly focus on the degree because the individual’s perceptual process is highly susceptible to slightly varying stimuli strengths (Trommsdorff 2008, p. 238) which may result in different perceptual outcomes.

Research on product availability investigates the effects of a low degree of prod- uct availability in particular and calls this occurrence scarcity (Verhallen 1982). This is a fundamental concept in traditional micro-economic theory (Lynn 1992b; Verhallen 1982). Approximately 48 years ago, the impact of scarcity on consumer behavior and perception was investigated from a psychological perspective (for example Brehm 1966; Brock 1968). This effect manifests itself in the so-called scarcity effect.

Verhallen and Robben (1995) identify different conditions under which people perceive products as scarce. These circumstances can be described as types of de- creased availability. Verhallen and Robben (1995) further distinguish between unavailability, restricted availability (a product is available only for certain peo- ple), conditional availability (a product is available only if certain criteria are met) and limited availability due to market or non-market circumstances. When talking about limited availability, the product is freely accessible but demand or supply- related circumstances lead to the fact that the product is perceived as scarce (Verhallen and Robben 1995). This thesis will focus on the supply-induced lim- ited availability due to market circumstance because it is most suitable to the sec- ond-hand C2C-context. The supply-related limited availability can be further dis- tinguished into quantity-related, time-related, and place-related or randomly gen- erated limited availability. Figure 3 provides an overview of possible types of lim- ited availability. The highlighted path in Figure 3 represents how limited product availability is understood in this thesis.

Figure 3: Types of Limited Availability.

illustration not visible in this excerpt

Source: Own Illustration Adapted from Winter 2009, p. 54.

In summary, in our study limited availability is defined by the available quantity (supplied) of a branded product in the second-hand market. Besides, the C2C- environment implies that the availability of products depends on the amount of second-hand products offered by private consumers.

2.6 Theories on Scarcity Effects

In the following, three popular theories regarding the effect of scarcity will be elucidated. These are commodity theory (Brock 1968), reactance theory (Brehm 1966) and scarcity as a heuristic cue (Cialdini 1993; Lynn 1989). Commodity the- -ry is presented because it has been widely employed to study the effects of scar- city (Lynn 1991; Verhallen and Robben 1995). Likewise, reactance is a theory that has been extensively studied in social psychology and in the context of scarcity (Clee and Wicklund 1980). Moreover, Verhallen and Robben (1995) present the scarcity heuristic in their review on scarcity theories and several journal articles on scarcity effects use the scarcity heuristic as a starting point (for example Ditto and Jemmott 1989; Gierl and Huettl 2010; Gierl, Plantsch, and Schweidler 2008; Lynn 1992b; Suri, Kohli, and Monroe 2007).

C ommodity theory (1968) states that consumers judge commodities according to their degree of availability. More precisely, scarce commodities have a higher perceived value than abundantly available ones (Worchel, Lee, and Adewole 1975). Brock (1968, p. 246) defines a commodity as anything that can be moved from an individual to individual and which is beneficial to its owner. The term ‘value’ is defined as the utility or the attractiveness and desirability of a commodi- ty (Brannon and Brock 1992; Lynn 1991). Hence, the evaluation of a commodity does not solely depend on the commodity’s functional and intrinsic properties; it is also affected by its supply- and demand-related characteristics (Brock 1968, p. 246). Several studies reveal (Lynn 1992b; Verhallen and Robben 1995) that scar- city which is provoked through market circumstances has a greater effect on the desirability of commodities than scarcity based on accidental or non-market in- duced reasons (Lynn 1992b; Verhallen and Robben 1995).

Reactance theory (Brehm 1966) starts from the premise that consumers usually feel they are free to choose between the goods on offer (Hammock and Brehm 1966, p. 546). The perception by a consumer that his/her freedom to possess a good is threatened by product scarcity evokes a psychological state - reactance - that motivates him to endeavor to retain his threatened freedom of choice (Brehm 1966, p. 15-16; Worchel 1992). Consequently, this perceived threat of losing the freedom of choice increases the desirability of scarce objects (Brehm 1966, p. 15).

Another approach is the more recent scarcity heuristic (Cialdini 1993; 2001; Lynn 1992a, 1992b), which argues that consumers judge a scarce product on the basis of the ‘scarce = attractive’ heuristic (Esch and Winter 2010, p. 24; Gierl and Huettl 2010). Product scarcity serves as a heuristic cue from which consumers di- rectly infer higher product value (attractiveness and desirability) (Cialdini 2001, p. 219; Lynn 1992a). According to Cialdini (2001, p. 228) consumers cannot cog- nitively refuse to apply the scarcity heuristic even if they are aware of its effect since it has a strong physical aspect that hinders consumer’s ability to think (Cialdini 2001, p. 228) and leads to automatic responses (Cialdini 2001, p. 8). Cialdini (2001, p. 219) claims that individuals have learnt through their socializa- tion as consumers that scarce products are better than abundant ones. Another cue-based approach is the ‘scarce = expensive’-heuristic (Lynn 1989; 1992b). It states that consumers infer a higher price from scarce products. Lynn (1992a) ar- gues that individuals have internalized this relationship from naïve economic the- ories. Moreover, Lynn (1992a) states that consumers want expensive products more than cheap ones. Through the higher price estimate (mediator) both the per- ceived quality and the perceived status of the product increase (Lynn 1989, 1992b). On the one hand, this can be explained through the price-quality relation- ship (for example Rao and Monroe 1989) according to which a consumer links higher prices to higher quality (Lynn 1992b). On the other hand, the product’s costliness might signal a higher social status of the person who possesses the scarce product. Figure 4 illustrates the relationships outlined above.

Figure 4: Model of the Scarce = Expensive Heuristic.

illustration not visible in this excerpt

Source: Own Illustration Adapted from Lynn 1992b, p.70.

2.7 Evaluation of Scarcity Theories regarding Applicability to Research Context

To the best of our knowledge, no one has ever stated that scarcity theories are only applicable to new products markets. Thus, we use their findings and transfer them to the second-hand market. However, we will be examining every theory individually in terms of its applicability to second-hand markets and products as well as the selected brand perceptions.

Commodity theory is valid for all marketable products (Lynn 1991). Since a used product can be resold, we regard commodity theory as also applicable to second- hand products. Another precondition of commodity theory is that there has to be a general interest in the product from the consumer’s point of view (Parker and Lehmann 2011; Verhallen 1982). Consequently, the findings of commodity theory should be evident when the degree of product involvement is rather high (Winter 2009, p. 57). Product involvement can be defined as a consumer’s stable percep- tion of the relevance of a certain product category based on inherent values, inter- ests and needs (Zaichkowsky 1985). In general, a durable product is said to be high-involving (Laurent and Kapferer 1985). This is due to the fact that if the pur- chase of a durable product turns out as a bad purchase, the consumer is stuck with a bad product for a longer time (Laurent and Kapferer 1985). Given the fact that second-hand products are (semi)-durables, we regard this theoretical premise as met and commodity theory as applicable to the second-hand market. In commodi- ty theory, perceived value, which can also be understood in terms of utility, is the main construct affected by product scarcity. Since the underlying definition of brand image also incorporates product utilities, the expectation of an influence of scarcity effects on brand image may be justifiable. The same holds true for the remaining brand perceptions, perceived quality and perceived brand prestige. Both perceptions comprise a functional (quality) and prestige (social) value (Sweeney and Soutar 2001) allowing for the transfer of findings from commodity theory to these perceptions.

A condition for applying reactance theory as specified by Worchel (1992) is that the suggested scarcity-value relationship only accounts for goods which consum- ers think they should have access to. It is only in cases where consumers feel that possessing that particular good is a critical freedom or right and scarcity signifies a threat to this right or freedom (Verhallen and Robben 1995) that this theory is applicable. However, consumer reactance is unlikely to be manifested in the case of the scarcity of an individual product from a specific product category that usu- ally comprises a variety of alternatives (Gierl and Huettl 2010). In such cases, consumers will probably not interpret the scarcity of one kind of product as a crit- ical restriction of their freedom of choice (Gierl and Huettl 2010). With regard to second-hand products, we propose that their possession is unlikely to be perceived as a critical right since consumers still have the freedom to acquire the brand new equivalent in the primary market. In conclusion, reactance theory does not seem particularly applicable to the second-hand market and therefore will not be further discussed.

The ‘scarce = expensive’ as well as the ‘scarce = attractive’-heuristics are more likely to be applied when other evaluative information on the product is not pre- sent (Jung and Kellaris 2004, Lynn 1989). For example, if product price infor- mation, quality information and brand name information are given, the activation process of a heuristic is weakened (Jung and Kellaris 2004). Nonetheless, we as- sume that since we will not give any exact information on the price of the second- hand products in our following empirical research (chapter 4 - 5) and there gener- ally exists uncertainty surrounding the quality of second-hand products, the condi- tion of ‘little other evaluative information’ is fulfilled. Thus, the scarcity heuris- tics may also be applied to second-hand products of low availability. Since the construct affected by the scarcity heuristic is again the perceived value of a prod- uct, we justify the transfer of the scarcity heuristic findings to the selected brand perceptions based on the same reasons for our use of commodity theory.

Table 3 indicates whether the theories in question are suited or not to our research context. It becomes evident that the commodity theory as well as the scarcity heuristics are applicable to predicting the potential effects of scarcity in the secondhand market and on brand perceptions.

Table 3: Assessment of Selected Theories on Scarcity Effects.

illustration not visible in this excerpt

After having evaluated the applicability of selected scarcity theories to our thesis, we will now briefly examine how product availability has been manipulated in previous studies related to supply-induced limited product availability. A consid- erable number of studies on scarcity distinguish between high (abundant) and/or low (scarce) product availability (condition) (for example Fromkin et al. 1971, Lynn 1989, Worchel, Lee, and Adewole 1975). Only Verhallen (1982) investi- gates the effects of a degree of product availability on product preference by em- ploying three levels of product availability - each level with an exact number of products available (low: 6/ medium: 16 / high: 30). For the purpose of our thesis, we will use a combination of the two manipulations and define second-hand prod- uct availability within the range of three levels (no/ low/ high), which has the ad- vantage of also considering the zero point from our product availability definition (more information on the operationalization can be found in the forthcoming chapter 4.4). This is important to bear in mind since our hypotheses will be pro- posing level-specific effects.

3. Development of Hypotheses

3.1 Main Effect

Mont, Dalhammar, and Jacobsson (2006) emphasize that it is important for companies competing on an established brand name to consider that associations with second-hand products may harm the company’s image. This view is supported by Clarke (2010, p. 240), who refers to brand strategists’ warnings that online second-hand markets for branded goods in particular contribute to the creation of an unfavorable brand identity, which is regarded as an antecedent of brand image (Heylen, Dawson, and Sampson 1995). Still, it remains unclear why brand image is possibly negatively affected by second-hand products.

One explanation for the postulated negative effects of branded second-hand prod- ucts on brand image in the context of the textile industry may be the existence of consumers who have negative associations with second-hand clothing and per- ceive these products as rubbish (Roux and Korchia 2006) or as mentioned in chapter 2.3 with little hygiene. These consumers believe that clothes can only be worn by a single owner and fear contamination (see chapter 2.3). Without explic- itly referring to clothing goods, this is supported by Brace-Govan and Binay (2010), who state that there is a only a thin line between ‘trash’ and ‘treasure’ when it comes to second-hand products. The second-hand phenomenon can be compared with that of counterfeit products (they share certain characteristics with second-hand products such as the lower quality and lower price) which are be- lieved to ‘cheapen the brand image’ of the corresponding genuine brand (Wilke and Zaichkowsky 1999).

When a brand is offered second-hand, it implies the coexistence of two different priced-versions of the same brand where the second-hand product’s price is lower compared to the new product. This similarity allows us to draw parallels to the area of price promotions, where the discounted product can be compared to the second-hand product. In this research context, there is empirical support for nega- tive spillover effects on brand image. Price promotions usually reduce the con- sumer’s reference price (Blattberg, Briesch, and Fox 1995; Hunt and Keaveney 1994), which can result in unfavorable brand evaluations and consumer confusion (DelVecchio, Henard, and Freling 2006; Jørgensen, Taboubi, and Zaccour 2003, Raghubir and Corfman 1999). A study by Montaner and Pina (2008) provides empirical evidence that frequent price promotions - a context comparable to that of offering second-hand branded products at a consistently lower price than their new equivalents - erode brand image. In addition, Yoo, Donthu, and Lee (2000) argue that frequent price promotions may communicate a low-quality brand im- age. Still, the transfer of these findings is only valid to some extent as research on price discounts refers only to new products.

Furthermore, the co-existence of two differently priced branded products may be perceived as a price inconsistency. This can lead to consumer confusion and con- flicting ‘value’ impressions, which results in unfavorable evaluations of brand im- age (Raghubir and Corfman 1999). In general, consistency is an important factor for a positive brand image (Farquhar 1989) and the literature suggests that price inconsistency negatively affects brand equity (Erdem and Swait 1998).

It thus seems reasonable to suppose that the direction of the general effect of sec- ond-hand product availability on brand image is negative (negative spillover ef- fect). However, the present study considers second-hand product availability in a relative rather than an absolute sense. Therefore, we additionally take the findings relating to a low degree of product availability into account (see chapter 2.5-2.7). As discussed in chapter 2.7, commodity theory (Brock 1968) as well as the scarci- ty heuristics (Cialdini 1993) may be applicable to second-hand products and ef- fects on brand perceptions such as brand image. This implies that if a branded second-hand product’s availability is limited, the perceived brand image may be affected positively. However, the negative spill-over effect from second-hand products per se needs to be considered as well. We suppose that the positive scar- city effect offsets the negative effect of second-hand products, adding up to a ze- ro-effect on brand image. In contrast to scarce second-hand products, abundantly available ones cannot benefit from this positive compensating effect (Worchel, Lee, and Adewole 1975) with the result that the negative effect of second-hand products may spill over onto brand image. In the following, the specific effects of a high degree of availability of second-hand products on brand image will be fur- ther discussed.

It is said that brand managers who wish to maintain a well-known brand’s reputa- tion, which can also be part of a brand’s image (Jacoby und Mazursky 1984), should be careful when employing deep price promotions (Drozdenko and Jensen 2005; Moore and Olshavsky 1989). The effect of deep price promotions can be compared with the effect of a deep price difference between the new and the sec- ond-hand product. This price difference is particularly pronounced when there is a high availability of second-hand products and is due to the fact that C2C second- hand market prices are mainly determined by supply and demand. In this context the law of supply states that the greater the supply, the lower the price (Dooley 1973, p. 157). Thus, the inferred second-hand price from the consumer’s point of view will be lower the more of the brand is available in the secondary market.

Attribution theory (Heider 1958; Kelley 1973) provides a more specific line of reasoning for the explanation of the possible effects of a high degree of second- hand product availability on brand image. This psychological theory suggests that consumers are likely to make causal inferences with reference to various events (Theotokis and Pramatari 2012). This process of inferring causes is termed at- tribution (Drozdenko and Jensen 2005). Attributions may influence consumers’ evaluation of aspects of the product (Kelley and Michela 1980) such as brand im- age. Moreover, attribution theory proposes that causal inferences are grounded in a consensus involving other consumers’ judgments. For example, high consensus among consumers is attributed to the brand (stimulus attribution) and low consen- sus is attributed to the consumer (Folkes and Kotsos 1986). If we apply attribution theory to our own object of investigation, we can conclude that consumers may make causal attributions to the fact that a particular brand is so highly on the sec- ond-hand market. The fact that a brand is highly available may be interpreted as indicating a high level of consensus among the consumers who resell the second- hand products in the C2C-market. It is therefore more likely that causal inferences will be attributed to the brand and have a spillover effect on brand perceptions. Backing this up, attribution theory is also applied in explaining the effects of price promotions (which we have identified as a ‘related’ research area) on brand eval- uations (for example Grewal et al. 1998; Low and Lichtenstein 1994). We further expect that when a brand is highly available in the second-hand market, the “dis- posed” character of second-hand products will be reinforced. The fact that con- sumers sell the branded product heavily may indicate that there is a strong con- sensus that the branded product has reached the end of its social lifetime (psycho- logical obsolescence). This implies that the branded product (respectively the brand) may not be desirable, popular or trendy anymore (attribution). The fact that these three attributes are rather unfavorable brand image dimensions (Jacoby and Mazursky 1984; Low and Lamb 2000) provides another argument for a nega- tive spillover effect of second-hand. This is supported by findings by Drozdenko and Jensen (2005) showing that especially low price promotions can evoke infer- ences or concerns that the product is outdated or old. Taking all the aspects elabo- rated above into account, we suggest:

Hypothesis 1 (H1): A high degree of availability of second-hand products results in a negative impact on brand image compared to the condition of no information on second-hand product availability.

3.2 Moderating Effect of Product Type

Baron and Kenny (1986) define a moderator variable as a third variable that changes the strength or direction of a relationship between an independent varia- ble and a dependent variable. There are some reasons to assume a moderating ef- fect of the product type in combination with second-hand product availability on specific brand perceptions. First, product characteristics are said to guide the way in which a consumer processes (brand) information (Suh 2009). Second, past studies (for example Gierl and Huettl 2010; Ku et al. 2013) have found interaction effects of product type and limited product availability on product evaluation. Whereas Gierl and Huettel (2010) distinguish between conspicuous and non- conspicuous products, Ku et al. (2013) classify products into hedonic and utilitari- an ones. The latter product categorization is commonly analyzed and applied in marketing literature (for example Dhar and Wertenbroch 2000; Hirschman and Holbrook 1982). Moreover, hedonic and utilitarian products are studied in price promotion literature (Montaner and de Chernatony 2011). Hence, in this thesis we will also concentrate on the hedonic versus utilitarian product classification.

Consumers’ product selections and assessments are driven by both hedonic and utilitarian considerations (Dhar and Wertenbroch 2000; Ku et al. 2013). Usually, the consumption of a lot of products involves both dimensions to a certain extent (Batra and Ahtola 1991), nevertheless consumers characterize some products as mainly hedonic and others as primarily utilitarian (O'Curry and Strahilevitz 2001; Okada 2005). Hedonic products are usually taken into consideration in terms of their potential for self-enhancement and self-expression (Mort and Rose 2004). Further, they are considered to be more personal, pleasurable, (Ahtola 1985) and value-expressive (Johar and Sirgy 1991). The consumption of a hedonic product is principally driven by emotional and social values (Mort and Rose 2004; Sloot, Verhoef, and Franses 2005). Typical examples are luxury watches or designer clothes (Dhar and Wertenbroch 2000). In contrast, utilitarian goods offer mainly functional, practical and performance-related benefits to the consumer (Batra and Ahotala 1991), and their consumption is therefore usually cognitive-driven and instrumental (Ryu, Park, and Feick 2006). Typical examples are microwaves and dishwashers (Sen and Lerman 2007; Strahilevitz and Myers 1998).

In utilitarian product groups, branded products are primarily differentiated by quality (Sloot, Verhoef, and Franses 2005) and how well they function (Leclerc, Schmitt, and Dubé 1994). So, when it comes to utilitarian products, consumers focus more on attributes such as reliability and durability (Prendergast, Chuen, and Phau 2002). This is plausible since utilitarian goods have longer life cycles than hedonic products (Ainslie, Drèze, and Zufryden 2005). Similar to the argu- mentation in chapter 3.1, we assume that utilitarian second-hand products being highly available could be interpreted as a strong consensus on the product’s at- tributes among the consumers who offer their products on the C2C-market. Con- sequently, the reason for this high availability may be attributed to the branded product. Due to the utilitarian nature of the product, one attribution may be that the branded second-hand product must be of good quality and durability because otherwise it could not be re-sold in the secondary market and would be fit for the scrap heap (the latter is excluded in our second-hand product definition). This positive attribution is assumed to spill over to the brand’s perceived quality. In this context the question arises as to why consumers would resell a utilitarian product of very good quality. This can be explained through the existence of s small consumer groups called innovators and early adopters who always want to have the latest models of products (Goldsmith and Flynn 1992).

In contrast, hedonic goods tend to be fashion goods (Verhagen and van Dolen 2011) that reach the end of their psychological lifetime more quickly compared to utilitarian products. Therefore, hedonic products are rather judged on their capability to perform in the short term (Phau, Sequeira, and Dix 2009) and quality and durability is not such a crucial issue as for utilitarian products.

In addition, the explanation of an attribution of high quality is plausible since it can be seen against the background of a practice called planned obsolescence, which applies to durable goods and is mostly present in the consumer electronics industry (Guiltinan 2009, Purohit 1992). More precisely, it is referred to physical obsolescence mechanisms such as ‘death dating’ or restricted functional life de- sign, which describe the practice of many electronics manufacturers of intention- ally incorporating predetermined breaking points in their (utilitarian) products (see here and in the following, Guiltinan 2009; N.N. 2013). This strategy reduces products’ usable lives and therefore their durability. Consequently, replacement purchasing by consumers is sped up and stimulates new product sales. However, since utilitarian products, such as backpacks (Buil, de Chernatony, and Montaner 2013), can also be non-electronic, it should be emphasized that the latter argument only holds true for ‘technical/electronic’ utilitarian products. Based on the above line of reasoning, our next hypothesis is as follows:

Hypothesis 2 (H2): Product type moderates the relationship between the degree of second-hand product availability and perceived quality. When the product is utilitarian, a high degree of second-hand product availability results in a positive impact on a brand’s perceived quality compared to the condition of a high degree of second-hand product availability and a hedonic product.

3.3 Moderating Effect of Price

Product price as a key element of the marketing mix (Yoo, Donthu, and Lee 2000) has been used in prior research on limited product availability (for example Lynn 1989). In line with the price operationalization of Suri, Kohli, and Monroe (2007), who did research on scarcity effects, we use a high and a low price as the price levels. This was done because the prices of new products and second-hand products are interrelated, which allows for more specific predictions on the effect of the degree of second-hand product availability.

Perceived brand prestige is often strongly tied to the fact that the use of a particu- lar brand is only given to an exclusive clientele (Vigneron and Johnson 1999). According to the ‘scarce = expensive’-heuristic, when the availability of second- hand products is limited, the assumed price and the status of the second-hand product increase. However, given a high availability of second-hand products, it is implied that there are significantly lower priced second-hand products available. As a result, consumers who were not able to afford the new branded product at its regular price would then be able to obtain the branded product (Porter and Sattler 1999; Purohit 1992). Scitovsky (1994) thus refers to the second-hand market as a market for poor people. As a result, second-hand products (similarly to counter- feits) are allowed free ride on the prestige benefits of brands which are tied to brand new products (Grossman and Shapiro 1988). Therefore, the perception of the brand’s prestige (which is affected by the second-hand product availability) may decrease since more people are able to acquire the branded product. Vigne- ron and Johnson (1999, p. 1) note that “if virtually everyone owns a particular brand it is by definition not prestigious”. However, we propose this negative ef- fect is likely to occur if the product has a high price (compared to a low price). This is due to the fact that for lower priced products the issue of the co-existence of lower priced second-hand products does not matter as much because most of the people have the financial means to buy an original low-priced product. More- over, the issue is most relevant when the second-hand product is highly available because then the price of the second-hand product will be lower (compared to a low availability) and even more people will be able to acquire the product. Thus, we predict:

Hypothesis 3 (H3): Price moderates the relationship between the availability of second-hand products and perceived prestige. When the product has a high price, a high degree of second-hand product availability results in a negative impact on the brand’s perceived prestige compared to the condition of a high degree of second-hand product availability and a low price.

3.4 Interactive Moderating Effects of Product Type and Price

We further propose that the negative effect on perceived prestige which is stated in hypothesis 3 will be reinforced depending on the product type. Hedonic prod- ucts are usually products that are tied to self-enhancement and symbolic aspects are said to be a subset of the hedonic dimension (Ryu and Park 2006). Hence, prestige plays a greater role in (high-priced) hedonic product categories (Verha- gen, Boter, and Adelaar 2010) due to the potential signaling value of hedonic products to other people (social groups). Consumers- when faced with utilitarian products - apply a more cognitive decision-making process, and focus on the product’s practical benefits and main tangible properties instead (Mano and Oliver 1993; Voss, Spangenberg, and Grohmann 2003). Thus, we surmise that a branded product’s prestige may become of secondary importance for utilitarian products. When there is a high availability of (lower priced) hedonic second-hand products, this might have a stronger negative effect on perceived prestige. This is because an essential aspect of the hedonic product (the status and self-enhancement value) is reduced due to the greater accessibility of these products.

Hypothesis 4 (H4): Product type moderates the relationship between the availa- bility of second-hand products, price and perceived prestige. When the product is hedonic and has a high price, a high degree of availability of second-hand products results in a stronger negative impact on the brand’s perceived prestige compared to the condition of a high degree of second-hand product availability, a high price and a utilitarian product.

We now argue for an additional interaction effect of price with regard to hypothe- sis 2. If there is only little product information present, price is regarded as an in- dicator of quality; this is known as the price-quality relationship. More precisely, consumers tend to rely on price as a cue to infer the product’s quality (for exam- ple Milgrom and Roberts 1986; Rao and Monroe 1989). Consequently, we assume that consumers infer a higher quality from a high-priced product compared to a low-priced one. Therefore, the quality implication of a high price may support the positive perceived quality attributions of a high degree of second-hand product availability for utilitarian products, which results in an additive effect. Thus, the effect assumed in hypothesis 3 is assumed to be strengthened through the moder- ating role of price. Therefore:

Hypothesis 5 (H5): Price moderates the relationship between the availability of second-hand products and product type and perceived quality. When the product has a high price and is a utilitarian product, a high degree of availability of second-hand products results in a stronger positive impact on the brand’s perceived quality compared to the condition of a high availability of second-hand products, a utilitarian product and a low price.

3.5 Control Variables

Continuous variables that are not an element of the main manipulation but can have an impact on the dependent variable(s) are called covariates (Field 2013, p. 479) or control variables (Rutherford 2011, p. 216). Thus, variables that potentially influence one or more of the thesis’ dependent variables are identified as potential control variables. We do not consider a simultaneous impact of the control variable on all three dependent variables as a necessary condition in a first step. However, in chapter 4.7 we will examine all covariates in detail and will further verify whether they qualify as control variables.

First, brand attitude which describes consumer’s overall negative or positive evaluation of a brand (Farquhar 1989) is identified as a potential control variable. Existing research provides statistical evidence that there is a positive relationship between brand attitude and brand image (Chang and Chieng 2006; Keller 1993). Brand familiarity, which concerns one's prior experiences with a given brand (Jamal and Goode 2001), represents the second control variable, since the litera- ture suggests that that brand image is positively influenced by brand familiarity (Low and Lamb 2000; Martínez and de Chernatony 2004). Additionally, we con- trol for brand sensitivity, which is the degree to which consumers attach im- portance to brand names and use it as a quality cue (D’Astous and Gargouri 2001) and identity signal. This variable is considered to be positively related to a brand’s perceived prestige since a brand sensitive consumer is more likely to perceive a well-known brand as prestigious (Jamal and Goode 2001). The concept of brand sensitivity is also strongly tied to prestige sensitivity, which will be another poten- tial control variable. Prestige sensitivity is defined as positive perceptions of the (high) price cue and based on ideas of status and prominence (Lichtenstein, Ridgway, and Netemeyer 1993).

3.6 Descriptive Variables

In order be able to later characterize our sample in more detail, several descriptive variables are included. First, attitude towards second-hand products is integrated since previous research demonstrates that consumers have differing attitudes of acceptance and refusal regarding second-hand products (see chapter 2.3). Second, frequency of second-hand shopping (Darley and Lim 1999) is included, which brings in a behavioral component. In addition, descriptive variables that are close- ly related to the moderator variables of this study, namely product type and price, and are linked to arguments in the hypotheses development are integrated. Thus, product knowledge and product involvement will be integrated. Product knowledge has been shown to predict if consumers are more likely to use heuris- tics (Suri, Long, and Monroe 2003) such as the scarcity or the price-quality heu- ristic. Product involvement on the other hand is a necessary condition for the ap- plication of the effects stated in commodity theory. As price is included as a mod- erator, the last descriptive variable is price consciousness, which brings in percep- tion of the price cue. It is defined as “the degree to which the consumer focuses exclusively on paying low prices” (Lichtenstein, Ridgway, and Netemeyer 1993, p. 235).

3.7 Conceptual Framework

All hypotheses (relationships) and variables are included in the final conceptual framework, which is presented in Figure 5.

Figure 5: Conceptual Framework.

illustration not visible in this excerpt

Note: Only the relationships of central interest are depicted in this figure.

4 Empirical Study

4.1 Pre-Test

A pre-test was conducted in order to determine a hedonic and a utilitarian product as well as a low and a high product price to be used in the main study. This approach follows other empirical studies that manipulated these variables (for product type: Sen and Lerman 2007; for price: Suri, Long, and Monroe 2003). Moreover, brand awareness and brand perceptions for certain brands were pre-tested in order to identify a suitable brand stimuli for the main study.

4.1.1 Stimuli Selection

For the product selection, secondary research was undertaken to determine typical utilitarian and hedonic goods that were used to represent each product type in past studies. Several other considerations influenced the selection of products to be included in the final study. The chosen products needed to be (semi-)durable and should actually be traded as second-hand products, for example on ebay.de, in or- der to increase the level of realism. Since the impact on brand perceptions was being assessed, it was important to choose product categories where brands play a role. This choice drew on previous research articles that also measure the impact on brand perceptions by using certain products. As a result, consumer electronics and clothing (Batra, Ahuvia, and Bagozzi 2012) as well as accessories and lug- gage (Desai and Keller 2009) were selected as the focal product categories. The latter three product categories are all part of the textile industry (Bundesverband des Deutschen Textileinzelhandels 2012). Furthermore, it was important to choose ‘gender-neutral’ products because gender differences can produce an additional source of variation. Since the thesis deals with branded second-hand products, we need later to choose a brand for each product type. In order to eliminate confound- ing effects and to be able to aggregate results across both products, we had to se- lect either one brand for both product types or two brands that are perceived as similar. Initially, we kept all options open and therefore took examples for both different product types that are offered by one brand in reality. As the same brand usually does not offer products in completely different industries, we needed to have some variation within each product type in terms of product category. Table

4 presents the selected products.

Table 4: Selected Products for the Pre-test.

illustration not visible in this excerpt

For each assumed product type, three products were chosen, resulting in a total number of six products to be pre-tested. All selected products met the previously listed criteria. However, some doubts remained about the appropriate product type for a smart phone since a phone was also identified as a utilitarian product in past studies (for example Sen and Lerman 2007) whereas we consider it as a hedonic product. We postulate that a smart phone comprises more ‘fun and hedonic fea- tures’ than a simple cellular phone. This is supported by prior research that has identified a mobile phone as being a highly conspicuous and symbolic product (Gierl and Huettl 2010), which can be positioned on the same level as hedonic products (Mugge, Schifferstein, and Schoormans 2006). Since the goal of the pre- test is to either confirm or refute these doubts, we integrated the smart phone into the hedonic category.

With regard to the criterion ‘traded on second-hand platforms’ there are differences between the products. This was revealed by a small informal, search analysis of ebay.de (results see Appendix A).³ The travel trolley and running shoes (1.193 items, 665 items) were offered to a far lesser degree than other products such as blue jeans (34.542 items). Nevertheless, all products are traded secondhand to some extent and were thus considered in the pre-test.

In order to identify different price points (low and high price) to be used in the final study, we followed an approach by Gardner (1971) and Gabor and Granger (1961). For each pre-tested product, the respondents had to indicate, “What is the lowest price at which you would still buy—the price below which you do not trust the quality?” and “What is the highest price you would be prepared to pay?” (Gardner 1971, p. 241). Thus, the respondents from the pre-test were asked to specify the two prices that represent their lower and upper price acceptability lim- its. A similar method to identify low and high prices for price manipulations has been employed by other researchers (for example Suri, Long, and Monroe 2003; Voss, Parasuraman, and Grewal 1998).

The selection of brands for the pre-test was driven by certain considerations. First, real brands were chosen instead of fictional ones. Real brands are richer in attrib- ute associations and their use in studies enhances ecological validity (Rodrigue and Biswas 2004). Furthermore, we intended to measure changes in prior brand perceptions and for a hypothetical brand there exist no prior brand perception un- less they are created.

[...]

---

¹ The terms ‘used’ and ‘second-hand’ will be employed synonymously in this thesis.

² The following journals were included in our search: Journal of Marketing, Journal of Marketing Research, Journal of Consumer Research, Marketing Science, Journal of Retailing, Journal of the Academy of Marketing Science, International Journal of Research in Marketing, Journal of Inter active Marketing, Journal of Service Research, Marketing Letters, Journal of Advertising Re- search, Journal of Consumer Behaviour, Advances in Consumer Research, Journal of Advertising ,International Journal of Retail and Distribution Management, International Review of Retail, Distribution and Consumer Research.

³ The analysis on ebay.de is only an informal ‘snapshot’ in terms of the number of used items of- fered for each product. In order to gain more exact numbers, an observation over a certain period of time would be more appropriate. However, our goal was only to achieve rough orientation.