The role of negative physical contagion in the online purchase of second-hand products

Abstract

The perception of negative physical contagion has been identified in the literature as a deterrent to the purchase of second-hand goods in an offline context. In view of the growth of second-hand shopping on the Internet, this article studies the effect of the salience of physical contact between the previous owner and the product in the purchase of second-hand products online. An initial experiment reveals a negative physical contagion effect on purchase intention, an effect mediated by perceived physical risk. However, many sellers of second-hand products on the Internet adopt presentation strategies aimed at emphasizing their similarity to new products. A second experiment shows that similarity with a new product reduces perceived physical risk for non-experts in the product category, while it decreases perceived quality for experts. In both experiments, the effect of disgust, which plays an important role in physical sales channels, is found to be not significant.

Keywords

contagion laws of sympathetic magic perceived quality perceived risk second-hand shopping online

Introduction

In 2016, the market for second-hand products in France amounted to €6 billion¹ and concerned 65% of French people.² This market is growing twice as fast as the market for new products,³ mainly due to the growth of online transactions, which accounted for 40% of second-hand purchases in 2017.⁴

Marketing research on second-hand shopping has focused mainly on physical distribution channels, identifying motivations (Roux and Guiot, 2008) and deterrents (Bezançon, 2012). These studies highlight in particular the important role that the contagion effect can play in second-hand shopping.

This effect was identified through the anthropological theory of the laws of sympathetic magic (Frazer, 1959 [1890]; Mauss, 1950 [1902]; Tylor, 1974 [1871]). According to the law of contagion, physical contact between an individual and a product gives rise to a permanent physical or symbolic transfer. Nemeroff and Rozin (1994) make a distinction, on the one hand, between physical contagion (transmission of residues or germs, e.g. microbes) and symbolic contagion (transmission of a particular characteristic, such as a personality trait), and on the other hand, between negative contagion and positive contagion. The type of contagion depends on the information available about the person who was in physical contact with the product. In the absence of information, perceived contagion is physical and negative (Argo et al., 2006; Di Muro and Noseworthy, 2013) and involves the belief that residues or germs can be transmitted simply by contact with the product (Nemeroff and Rozin, 1994). Various authors have demonstrated the role of negative physical contagion in offline second-hand shopping (Dehling, 2013; Gregson and Crewe, 2003; O’Reilly et al., 1984; Roux and Korchia, 2006). The negative physical contagion effect is all the stronger if the product has a high degree of bodily intimacy, that is, if it comes into prolonged contact with the body (Dehling, 2013). Examples include clothing, shoes, jewellery and accessories, toys, articles used for childcare and, to a lesser extent, furniture and books. The more apparent it is that the product has already been handled (presentation, signs of wear, residues), the stronger the perceived negative physical contagion (Argo et al., 2006). In an offline context, its effects are based on feelings of disgust (Argo et al., 2006; Di Muro and Noseworthy, 2013; Roux, 2004a).

In the case of online purchases, however, potential buyers are not in direct contact with the product, but are able to see one or more photos before making their purchase decision (Abdul-Ghani et al., 2011), which makes it more difficult to determine whether the product has already been used or handled. In addition, the context of online shopping limits sensory stimulation (Helme-Guizon, 2001) and seems less likely to induce feelings of disgust. It is therefore interesting to study the existence of a negative physical contagion effect for second-hand purchases online, especially as many intimate products are now bought second-hand on the Internet.⁵

At the same time, observation of the practices of online second-hand sales websites, especially for professional sellers, reveals various presentation strategies for reducing the perception of contact between the product and the previous owner. The main photos of the products are often cropped (e.g. clothing on VestiaireCollective or Videdressing) or taken in a neutral environment, without the owner or his or her place of residence being visible. In addition, the products may be presented as new, variously in their original packaging (e.g. phones on PriceMinister), repackaged so as to look like the packaging of new products (e.g. video games sold in blisters on Micromania), or with a label put back into place and visible (e.g. clothing on Leboncoin). For websites selling both new and used products, these can be presented jointly, with a generic visual (e.g. CDs, DVDs, books and video games on the Fnac website). These presentation strategies, which involve significant efforts on the part of sellers, make second-hand products look like new products, while explicitly mentioning that they are indeed second-hand. Yet according to the second law of sympathetic magic – the law of similarity – a superficial resemblance between two entities indicates a fundamental resemblance between them, and they are perceived as identical. Resemblance to a new product could thus result in second-hand products being perceived more like new products and thus limit the effects of the law of contagion (Rozin, 1994). We can therefore legitimately wonder about the impact of these strategies, in a context of potential perception of negative physical contagion.

The research questions addressed in this article are as follows. Is a negative physical contagion effect observed when second-hand products are sold online? If so, by means of what mechanisms? And does the similarity of presentation of the second-hand product with a new product reduce the negative contagion effect?

A literature review first explains the concept of negative physical contagion. The potential effects on perceived risk (a key variable in second-hand shopping) and purchase intention are then presented. The effect of resemblance to a new product on perceived risk, as well as on the perceived quality of the product, is then explored. An initial experimental study reveals a negative physical contagion effect of second-hand products sold online on perceived physical risk and purchase intention. A second experimental study reveals the influence of similarity to a new product on the negative physical contagion effect of second-hand products sold online: similarity reduces perceived physical risk for non-experts in the product category, while it adversely affects perceived quality for experts. As well as their theoretical contributions, these findings also have implications – discussed in the final part of the article – for online sellers of second-hand products.

Literature review

The concept of negative physical contagion

The concept of negative physical contagion is very much in vogue in marketing research (Huang et al., 2017). An extensive literature shows that perceived negative physical contagion affects the process of purchasing many types of products: second-hand products (Dehling, 2013; O’Reilly et al., 1984; Roux and Korchia, 2006), shared (Bardhi and Eckhardt, 2012) or remanufactured products (Abbey et al., 2015), and new products that have been tried out in stores (Argo et al., 2006) or whose packaging has deteriorated (White et al., 2016). Surprisingly, these studies only concern offline shopping: to our knowledge, no-one has looked at the effects of negative physical contagion associated with products sold through the Internet.

In these studies, negative physical contagion is not measured directly, but its effects are ascertained through its determinant, namely perceived contact between the object and an individual. The authors of these studies vary the salience of contact, defined as whatever makes it more obvious that the product has been used (Argo et al., 2006). Contact with the item does not have to be actually seen, just that it be inferred by the potential buyer. For example, a product’s negative physical contagion may result from its dirtiness (O’Reilly et al., 1984), from the fact that it has been damaged (Di Muro and Noseworthy, 2013; White et al., 2016), its odour (Bardhi and Eckhardt, 2012) or its having been unfolded or presented untidily (Castro et al., 2013; Reynolds-McIlnay et al., 2017). Similarly, some product categories make it obvious that the product has already been used and give rise to the perception of negative physical contagion: this is the case for second-hand products (Ackerman and Hu, 2017; O’Reilly et al., 1984; Roux and Korchia, 2006).

The marketing literature has shown that, in an offline context, the negative physical contagion effect is based on the emotion of disgust. For example, Argo et al. (2006) and Di Muro and Noseworthy (2013) show that disgust is a mediator of the negative influence of contact between the product and an unknown consumer on the intention to purchase or use it. With regard to second-hand purchases, Roux (2004a) shows that sensitivity to interpersonal disgust – defined as the tendency of some people to experience feelings of disgust when in direct contact with unknown, strange, sick or immoral individuals (Rozin et al., 2000) – is linked to a refusal to buy things second-hand for fear of the effects of contagion.

Negative physical contagion and similarity to a product that has never been used

Sometimes, a product does not give rise to a negative physical contagion effect even though it is obvious that contact with it has already occurred. This is particularly the case if the product bears a strong resemblance to a product that has never been touched (Rozin, 1994). In the case of second-hand buying, Roux (2004a) finds that some sellers present their second-hand products as new, a practice that reduces the perception of negative physical contagion.

Such absence of contagion is explained by the law of similarity, which is also derived from the laws of sympathetic magic (Frazer, 1959 [1890]; Mauss, 1950 [1902]; Tylor, 1974 [1871]). According to Rozin and Nemeroff (2002), the law of similarity is a psychological heuristic that leads the individual to evaluate an entity according to its superficial characteristics when these are identical to the characteristics of a known entity. These authors note that the effects of the law of contagion and of the law of similarity are opposed to each other. The law of contagion leads the individual to evaluate a product according to its past history, even if this history is unknown. The product concerned is thus different from what it seems to be: the perception that the product has been handled, even if the contact has left no visible trace, freights it with properties. Conversely, the law of similarity leads the individual to evaluate a product according to its superficial characteristics. The product is what it seems to be, and its apparent properties allow its fundamental properties to be grasped. In particular, strong resemblance to a product that has never been touched would suggest that no contact has occurred. Various psychological studies, particularly with regard to food, show that the effects of the law of similarity can reduce or even eliminate the effects of the law of contagion (Rozin, 1994; Rozin and Nemeroff, 2002).

The specificities of online shopping and the impact on perceived negative physical contagion from second-hand products

Work on the negative physical contagion effect offline emphasizes the roles of contact salience, disgust, and similarity to an untouched object. Analysis of the specificities of buying second-hand products online (presented in detail in Appendix 1) leads us to consider the role of these characteristics on the Internet.

First of all, we find that online second-hand shopping channels rarely provide personal information about who the products belonged to (Pantoja et al., 2016). The information available is generally limited to a pseudonym, a location and possibly an indicator of online reputation (Ghose, 2009; Li et al., 2009). According to Argo et al. (2006), this purchasing context is therefore conducive to a negative physical contagion effect associated with the products.

Second-hand products sold online may present a high contact salience. It certainly is impossible to inspect the product and see the various traces left by its past use: its presentation is limited to a few photos and some description (Abdul-Ghani et al., 2011). Note, however, that some aspects of photos may make contact with the owner more salient: signs of use may be visible – for example, a book with a damaged cover or a name written on it (Kim, 2017) – or the product may be shown in the photo in direct contact with the previous owner – for example, a mobile phone held in the hand (Doleac and Stein, 2012; Kapitan and Bhargave, 2013). Moreover, it has been shown in the context of offline shopping that the product description could make contact more salient – for example, the mere mention that the product is ‘good as new’ underlines the fact that it has been used (Ackerman and Hu, 2017).

The context of online second-hand shopping seems, on the other hand, less likely to give rise to feelings of disgust. Offline, second-hand products can be seen, touched or smelled in a stimulating sensory environment that often includes direct interaction with a salesperson or other consumers (Abdul-Ghani et al., 2011; Sherry, 1990; Solomon, 2008). But online, sensory stimulation is reduced: the item cannot be touched or smelled, and its perception is necessarily mediated by a screen (Helme-Guizon, 2001). Sellers and buyers have no face-to-face contact (Rafaeli and Noy, 2002; Yen and Lu, 2008), and are not physically present in the same place and at the same time (Chakravarti et al., 2002; Cui et al., 2008). Sensations and feelings are limited (Helme-Guizon, 2001) and emotional mechanisms are likely to play a lesser role, especially as purchasing second-hand products online calls for increased cognitive efforts. Indeed, the large number of items for sale online (Cameron and Galloway, 2005; Ghose et al., 2006; Singh et al., 2014) and the information available on products and sellers (Ghose, 2009; Gregg and Walczak, 2008; Li et al., 2009) make decision-making more complex (Ariely and Simonson, 2003).

Finally, the similarity of second-hand products to new products appears to be more pronounced online than offline, a fact that may limit the perception of products’ negative physical contagion because of the similarity law. The second-hand product offering format is standardized and in some respects identical to that of new products (Gregg and Walczak, 2008). While offline second-hand channels are often distinct from those for new products (Guiot and Roux, 2010), the boundary between new and second-hand products may be blurred on the Internet and many websites offer both (Ertz, 2015; Kauffman and Wood, 2006; Walczak et al., 2006).

In order to study a potential negative physical contagion effect in more detail, we now consider the variables that affect an online second-hand purchase.

The determining variables of second-hand purchases online

Marketing research on the negative physical contagion effect shows that it influences the variables determining a purchase (Argo et al., 2006). In addition to purchasing intent, which is considered a predictor of consumer behaviour (Kalwani and Silk, 1982), the literature emphasizes the role of perceived risk. Furthermore, in cases where the second-hand product looks like a new product, perceived quality – another key variable for second-hand purchases – may also be play a role.

Numerous studies on online shopping and second-hand shopping reveal the importance of perceived risk (Donthu and Garcia, 1999; Gabbott, 1991; Roux, 2004b; Tan, 1999). Defined as the perception of uncertainty about the potential negative consequences associated with the purchase (Bauer, 1960; Volle, 1995), perceived risk has six dimensions (Jacoby and Kaplan, 1972; Kaplan et al., 1974; Roselius, 1971). Of these, performance risk and financial risk have been studied most because of their high explanatory power for consumer behaviour, both offline (Agarwal and Teas, 2001; Suwelack et al., 2011) and online (Biswas and Biswas, 2004; Chen and Dubinsky, 2003; Forsythe and Shi, 2003). Performance risk concerns the fear that the product will not function as it is supposed to; and financial risk the fear that the product will need to be replaced or repaired (Jacoby and Kaplan, 1972; Stone and Grönhaug, 1993). Perceived financial and performance risks are exacerbated when purchasing second-hand online (Ertz, 2015; Lai et al., 2008). The literature also emphasizes the role of perceived physical risk, particularly for second-hand products with a high degree of physical intimacy (Dehling, 2013; O’Reilly et al., 1984; Roux and Korchia, 2006). Perceived physical risk is the fear that the product will endanger the user’s health (Jacoby and Kaplan, 1972; Stone and Grönhaug, 1993).

Research on second-hand shopping online also highlights the role of perceived quality (Ghose, 2009; Gregg and Walczak, 2008), defined as the consumer’s overall opinion regarding the excellence or superiority of the product (Zeithaml, 1988). Perceived quality is determining in second-hand shopping because of the difficulty of assessing it. Indeed, the second-hand market is characterized by the asymmetry of information between sellers and buyers, both offline (Akerlof, 1970) and online (Ghose, 2009). Sellers are more familiar with the quality of their goods than buyers, who are often unable to verify it. As a result, buyers tend to rely on quality indicators to assess the good before purchase. But the resemblance of the second-hand product to a new product could be treated as information provided by the seller on the quality of the product. Taking this variable into account thus allows us to study the overall effects of a strategy whereby the resemblance of the second-hand product to a new product is emphasized.

The literature review thus raises three research questions. While contact salience is plausible in the case of an online second-hand product offering, the feeling of disgust that explains the effects of negative physical contagion offline seems unlikely when buying on the Internet. So does perceived negative physical contagion influence second-hand purchase online? If so, by what mechanisms? The literature also underlines the possible strong resemblance between a second-hand product sold on the Internet and a new product. Does the law of similarity then reduce the negative physical contagion effect of the second-hand item? Two experimental studies allow us to answer these questions, taking into account the key variables of online second-hand shopping, namely perceived risk and perceived quality.

Study 1: Does the negative physical contagion effect exist online?

This first study has two objectives: to test the potential effect of negative physical contagion of second-hand products sold online on perceived risk and purchase intention; and to examine the mechanism underlying this effect. The effect of negative physical contagion is grasped in this study through its determinant, via contact salience. A moderating variable – sensitivity to interpersonal disgust – allows us to study the underlying mechanism. The conceptual model tested is presented in Figure 1.

Hypotheses

Higher contact salience gives rise to higher perceived negative physical contagion (Argo et al., 2006). If there is a negative physical contagion effect, a higher salience will result in a higher perceived physical risk: consumers are wary of the effect on their health of buying second-hand goods. In addition, perception of a negative physical contagion from the product may lead the consumer to stop using it (Dehling, 2013; Roux and Korchia, 2006). High contact salience is therefore also likely to lead to an expectation of unsatisfactory performance by the second-hand product and thus a potentially wasted financial outlay. We thus formulate the following hypothesis:

H1. Contact salience positively influences perceived (a) physical, (b) performance and (c) financial risk.

The literature highlights the influence of perceived physical, performance and financial risk in second-hand shopping (Dehling, 2013; Roux, 2004b). Perceived risk has also been shown to negatively influence purchase intention, in both offline (Grewal et al., 1998) and online contexts (Van der Heijden and Verhagen, 2004; Kwon and Lennon, 2009). If contact salience has a negative influence on the perceived risks, then these risks should have a negative influence on purchase intention. We thus formulate the following hypothesis:

H2. The influence of contact salience on purchase intention is mediated by perceived (a) physical, (b) performance and (c) financial risk.

In order to study the mechanism of the potential effect of negative online physical contagion, we draw on the concept of sensitivity to interpersonal disgust, which is a dimension of the personality trait of sensitivity to disgust (Haidt et al., 1994) and which has been used extensively in psychology and marketing (Goukens et al., 2007; Hodson and Costello, 2007; Kapitan and Bhargave, 2013; Olatunji et al., 2007). In an online context that is not very conducive to emotion, this concept has an advantage over a direct measure of disgust, since it allows observation of a possible negative physical contagion effect that might be restricted to individuals who are very sensitive to interpersonal disgust.

We hypothesize that individuals with high interpersonal disgust sensitivity are more likely to feel disgusted when viewing online second-hand product offerings, particularly if contact salience is high. According to Chaudhuri (1997), negative emotion can be considered as experiential information. Since risk is affected by the information available (Bauer, 1960), negative emotion may therefore increase the perception of risk (Chaudhuri, 1997, 1998). So we hypothesize that, for these individuals, perceived physical risk is higher, which in turn negatively influences purchase intention.

H3. Sensitivity to interpersonal disgust moderates the mediation of the influence of contact salience on purchase intention through perceived physical risk.

Experimental procedure

Two conditions were specified for this study: weak versus strong salience of contact with the product. In order to vary contact salience without modifying the objective quality of the product, we used the visibility in the photograph of contact between the owner and the product. The stimuli used were advertisements of second-hand products, in order to recreate the world of second-hand shopping online. However, to avoid attitude bias towards an existing website, we used a dummy site: ‘www.produits-occasion.com’. Since perception of negative physical contagion in an online purchase context may be difficult to detect, for stimuli we selected a product whose level of bodily intimacy is high and therefore likely to generate a strong feeling of physical contagion (Dehling, 2013). An initial pre-test conducted with 35 students – who evaluated the level of bodily intimacy of eight different products by giving their degree of agreement (from 1 to 7) to the statement ‘This product is often in contact with the body’ (M_book = 4.76; M_bag = 3.12; M_trainers = 6.20; M_sweatshirt = 5.34; M_desk = 4.64; M_bed = 6.07; M_phone = 4.44 and M_headphone = 5.45) – led us to select running shoes (hereafter ‘trainers’) for this study. The chosen model is a pair of mid-range trainers that may be suitable for more or less experienced runners, both men and women. To validate the choice of the model, a second pre-test was conducted with 63 students. The degree of agreement with the statement ‘These trainers could suit my needs’ is no different between respondents who run and those who do not (M_practice = 4.03; M_nonpratice = 3.73; p = 0.50) or between men and women (M_men = 3.90; M_women = 3.84; p = 0.89). Similarly, evaluation of the statement ‘These trainers have the technical characteristics I am looking for’ does not differ according to those who run and those who do not (M_practice = 3.90; M_nonpractice = 4.06; p = 0.67) and gender (M_men = 3.77; M_women = 4.19; p = 0.27). Finally, the difference in liking the colour of the trainers according to the gender of the respondent is not significant (M_men = 3.10; M_women = 2.53; p = 0.18). The same trainer was therefore used (size 41), photographed unworn for the condition of weak salience, and worn by a woman or man for the condition of strong salience. The high price of this product new makes purchasing it second-hand appropriate. However, since trainers may be stretched out of shape by being worn and may lose their cushioning qualities with prolonged use, we specified in the ad that the trainers had been ‘little worn, in good condition’. The stimuli are shown in Figure 2.

The 88 students (M_age = 23.44; 53.3% women) participating in this study, remunerated in the form of course credits, answered a questionnaire through Qualtrics software in an experimental laboratory. Respondents were randomly assigned to one or other of two conditions (inter-subject design). The questionnaire began with a general presentation of the subject of the study: second-hand shopping online. Respondents were then exposed to a purchase scenario (Appendix 2) followed by an advertisement. They were first asked about their purchase intention, and then about the perceived physical, performance and financial risks associated with buying the product. The concluding questions pertained to the control variables, the manipulation check, sensitivity to interpersonal disgust and the respondent’s general profile.

Measurement instruments

Verification of the manipulation of contact salience was carried out using an ad hoc item: ‘I realize that these trainers have already been worn’. The measure of purchase intent was derived from the work of Argo et al. (2006) and consists of a single translated item. For perceived risk, we used Stone and Grönhaug’s (1993) measure, which has the advantage of offering three items per dimension and can be adapted to the product concerned (Ayadi, 2010; Keh and Pang, 2010; Wiedmann et al., 2011). These items were subjected to a process of translation/back-translation, then adapted to the context of second-hand purchase (α = 0.703 for perceived physical risk, α = 0.584⁶ for perceived performance risk, α = 0.627 for perceived financial risk). The measure of sensitivity to interpersonal disgust is taken from Haidt et al. (1994). The most recent version of this measure, reviewed by Olatunji et al. (2007), contains only three items, which have been validated in a French context (Gil et al., 2009). Nevertheless, a pre-test conducted with 44 students underlined the low reliability of this three-item measure (α = 0.545). However, inclusion of three additional items from the second version of the scale (Haidt et al., 2002) resulted in better reliability (α = 0.773). In addition, the unidimensionality of the scale was verified through a principal component analysis and brings out a single factor embodying the six items of sensitivity to interpersonal disgust. We include two control variables in this experiment. Controlling for sustained involvement helps to avoid a response bias due to lack of interest in the product. Sustained involvement is measured using Strazzieri’s (1994) relevance-interest-attraction (RIA) scale. To simplify the questionnaire, however, we use a reduced version of the original scale with only three items (Merle, 2007; Rieunier, 2000) (α = 0.897). We also control for the respondent’s gender because our stimuli are adapted to this. The items of the measurement scales (all 7-point) are presented in Appendix 3.

Findings

We verified our manipulation by carrying out an analysis of variance (ANOVA) with the manipulation check item as a dependent variable and the contact salience condition as an independent variable. The result of the test is significant: contact salience is higher under the strong salience condition (M = 5.81) than under the weak salience condition (M = 4.98; F (1.86) = 6.08; p = 0.02). In addition, we verified that the product is liked as much under the strong salience condition (M = 4.02) as under the weak salience condition (M = 3.76; F (1.86) = 0.53; p = 0.47). Using the general linear model technique, hypothesis H1 was tested successively with perceived physical, performance and financial risks as dependent variables, contact salience as an independent variable, and gender and involvement as covariables. The result of the test is significant for perceived physical risk, which is higher under the strong salience condition (M = 4.98) than under the weak salience condition (M = 4.38; F (1.84) = 3.970; p = 0.049). However, the test is not significant for performance and financial risks (F (1.84) = 1.73; p = 0.19 and F (1.84) = 0.04; p = 0.84, respectively). There is no significant difference in scores depending on whether the salience condition is strong (M = 5.00 and M = 4.42, respectively) or weak (M = 4.66 and M = 4.47). Hypothesis H1a is confirmed and hypotheses H1b and H1c are disconfirmed.

All H2 and H3 hypotheses were tested simultaneously by means of the Preacher and Hayes (2008) PROCESS macro. Model 7 allows us to include three mediators acting in parallel, a moderator (on the effect of contact salience on perceived risk) and the two control variables, with 5000 samples generated by bootstrap (averages, correlations and distribution parameters of quantitative variables⁷ are shown in Appendix 4).

To test H2a, we follow the steps recommended by Baron and Kenny (1986) and then reviewed successively by Zhao et al. (2010) and Hayes (2013, 2018). We thus first look at the indirect effect of contact salience on purchase intention, via perceived physical risk. The effect of perceived physical risk on purchase intention is not significant (b = −0.25; t = −1.73; p = 0.08),⁸ but we clearly observe a negative indirect effect of contact salience on purchase intention through perceived physical risk (a × b = −0.16) with a 95% confidence interval excluding 0 [−0.56, −0.01]. Even in the absence of a significant relationship between the hypothesized mediator (perceived physical risk) and the dependent variable (purchase intention), the mediation is significant (Hayes, 2018: 115). In order to determine more precisely the type of mediation, we examine the direct effect of contact salience on purchase intention, which is not significant (c = 0.02; t = 0.05; p = 0.96). We can therefore conclude that mediation is ‘indirect only’ (Zhao et al., 2010). Hypothesis H2a is confirmed.

The same methodological approach is used to test the assumed mediating roles of perceived performance risk (H2b) and financial risk (H2c). In both cases, the indirect effects are not significant (a × b = 0.07 and a × b = 0.01, respectively) with 95% confidence intervals including 0 ([−0.03, 0.36]; [−0.07, 0.16]), which leads us to reject these hypotheses.

Testing hypothesis H3 shows that people’s level of sensitivity to interpersonal disgust does not alter the indirect effect of contact salience on purchase intent through perceived physical risk. Indeed, the interaction between contact salience and sensitivity to interpersonal disgust is not significant (b = –0.09; t = –0.42; p = 0.67), the Hayes (2018) moderated mediation index is almost zero (I = –0.03) and the associated confidence interval includes the value 0 [−0.05, 0.17]. Hypothesis H3 is rejected.

In total, the model corresponding to hypotheses H2 and H3 is able to explain 21.5% of the variance of purchase intention.

Discussion

The confirmation of hypothesis H1a shows that contact salience increases perceived physical risk and thus fears about the product’s hygiene and the transmission of germs or residues. This finding reveals for the first time a negative physical contagion effect in the context of online shopping and answers our first research question. Nevertheless, the lack of effect on perceived performance and financial risks suggests that perceived negative physical contagion is not large enough to deter use of the product and the financial outlay involved.

The confirmation of hypothesis H2a shows that perceived physical risk influences purchase intention and mediates the effect of salience on purchase intention. This finding confirms the value of studying the phenomenon of online negative physical contagion. However, unlike studies on offline shopping (Argo et al., 2006; Morales and Fitzsimmons, 2007), our results do not show a direct effect of contact salience on purchase intention. The absence of a direct link may be explained by the existence of competing mediation (Hayes, 2018), that is to say, another mediator whose effect is opposed to that of perceived physical risk. For example, we can think of embodied mental simulation, defined as a visual representation that allows the consumer to project him/herself into the process of using the product (Elder and Krishna, 2012). It has been shown that embodied mental simulation can increase purchase intention for a product that has already been used (Kim, 2017). Contact salience might stimulate embodied mental simulation, which could be a positive mediating effect. Finally, the rejection of hypothesis H3 suggests that interpersonal sensitivity to disgust does not play a part in the negative physical contagion effect online, whereas it has been shown to do so offline (Roux, 2004a). With regard to our second research question about the mechanisms of the negative physical contagion effect online, this absence of moderation and the mediating role of perceived physical risk, underline the possibility of a mechanism that is more cognitive than emotional. A summary of the results is provided in Appendix 5.

Study 2: Does similarity to a new product reduce the effect of negative physical contagion online?

Our third research question concerns presentation practices regarding second-hand and new products online (Ertz, 2015; Kauffman and Wood, 2006; Walczak et al., 2006). In Study 2, we examine to what extent similarity with a new product reduces the effect of negative physical contagion in second-hand shopping on the Internet, by acting on perceived physical risk and purchase intention. We also test the effect of similarity on perceived quality, so as to better understand the consequences of this presentation strategy by sellers. Two moderating variables are taken into account: expertise in the product category, which may alter the effect of similarity, and sensitivity to interpersonal disgust, which allows us once again to consider the mechanism of contagion. Figure 3 presents the conceptual model of this study.

Hypotheses

According to the law of similarity, a second-hand product that strongly resembles a new product should give rise to a similarity heuristic and should be perceived by the consumer as new (Rozin, 1994; Rozin and Nemeroff, 2002). Since a new product has in principle not been used, the main determinant of negative physical contagion – the perception of contact – would be absent and the negative physical contagion effect would no longer occur. A second-hand product that looks like a new product should therefore result in lower perceived physical risk.

We hypothesize, however, that this effect will differ according to the consumer’s expertise, since the level of expertise in the product category influences information processing (Alba and Hutchinson, 1987). Non-expert consumers have limited information processing capabilities, do not process information exhaustively, and use evaluation heuristics (Bettman and Park, 1980). When the product has multiple attributes, non-expert consumers tend to base their initial impression on those that are most visible (Maheswaran, 1994). A strong resemblance to a new product should therefore limit their perception of physical risk due to the heuristic of similarity. In contrast, expert consumers can process a large amount of information in depth (Maheswaran and Sternthal, 1990) and are able to evaluate each attribute separately (Raju et al., 1995). Their level of perceived physical risk should therefore not be changed by a superficial resemblance to a new item. We thus formulate the following hypothesis:

H4a. The effect of similarity on perceived physical risk is moderated by the level of expertise: similarity negatively influences perceived physical risk for non-experts; this effect is not observed for experts.

Resemblance of the second-hand product to a new product could also influence its perceived quality. Indeed, the consumer often uses heuristics to evaluate the quality of products, based on various attributes, such as price, packaging or the brand (Gerstner, 1985; Maheswaran et al., 1992; Orth et al., 2014; Rao and Monroe, 1989). Because product quality is particularly uncertain in the case of a second-hand purchase online (Ghose, 2009), presentation of the product as new could lead to a heuristic of quality evaluation. The use of heuristics to assess quality, however, depends on the level of expertise (Srivastava and Mitra, 1998). We thus formulate the following hypothesis:

H4b. The effect of similarity on perceived quality is moderated by the level of expertise: similarity positively influences perceived quality for non-experts; this effect is not observed for experts.

For the same reasons as in Study 1, we hypothesize that perceived physical risk influences purchase intention. We also again test the effect of sensitivity to interpersonal disgust, on a different product. We assume that when viewing an ad for a second-hand product sold online, individuals with strong sensitivity to interpersonal disgust experience an emotion of disgust and therefore perceive a strong physical risk. For non-expert individuals, the effect should be less pronounced if the product is presented as new. Accordingly we derive the following hypothesis:

H5a. Perceived physical risk mediates the effect of similarity with a new product on purchase intention. This mediation is moderated by perceived expertise and sensitivity to interpersonal disgust.

Many studies have shown the positive influence of perceived quality on purchase intention (Chaudhuri, 2002; Zeithaml, 1988). Consequently, we hypothesize that:

H5b. Perceived quality mediates the effect of similarity with a new product on purchase intention. This mediation is moderated by perceived expertise.

To ensure that the effects observed are due to the manipulations carried out, we include various control variables: on the one hand, sustained involvement, for the same reasons as in Study 1, and on the other, perceived financial risk and perceived performance risk (which were found to be non-significant in Study 1).

Experimental procedure

Two conditions were specified for Study 2: weak versus strong similarity with a new product. In order to vary this similarity, we use the absence versus presence of packaging around the product. Packaging is a characteristic attribute of new products (Warnier, 1999; Wever and Del Castillo, 2006), but is also used by some companies to sell second-hand products (Roux, 2004b). In the interest of external validity, in this study we work with another product whose level of bodily intimacy is high according to our pre-test: a headphone. This product differs from trainers by virtue of its technicality and has the advantage of being unisex and sold new in packaging. In the condition of weak similarity, the headphone is photographed without any packaging. In the condition of strong similarity, it is photographed with plastic packaging. We made sure not to use the original packaging, as this could have suggested that the product had never been unpacked. In addition, the presence of the original packaging might indicate that the owner is particularly careful and thus limit the internal validity of the study. Finally, the use of original packaging requires that it has been kept, something that is unusual and would make the results of the study difficult to generalize. The headphone’s original box is not shown, nor is its plastic shell. The packaging used for the stimuli does not carry a brand name associated with headphones. Only the PELD logo is visible, indicating that the item has plastic components. To ensure that respondents are informed about the second-hand status of the product, under both experimental conditions, the scenario explicitly mentions that the product has been used. A pre-test allowed us to verify this manipulation: 32 students exposed to the condition with packaging gave their degree of agreement on a 7-point scale to the assertions ‘this helmet has been used’ and ‘this helmet has been repackaged with plastic wrapping which is not its original packaging’. The averages are significantly greater than 4 (M = 5.09; p = 0.00 and M = 5.03; p = 0.01), emphasizing agreement. Once again, the product is featured in an online sales ad (Figure 4).

A total of 130 students (M_age = 22.40; 62.3% women), remunerated in course credits, participated in this study and answered a questionnaire via Qualtrics software in the experimental laboratory. The survey was conducted in a similar manner to that of Study 1.

Measurement instruments

Verification of the manipulation of similarity with a new product was carried out using an ad hoc item: ‘In the photo, the headphone looks new’. Measures of purchase intention, perceived physical risk (α = 0.886), performance risk (α = 0.701), financial risk (α = 0.868), sensitivity to interpersonal disgust (α = 0.767), and sustained involvement (α = 0.809) were the same as in Study 1. Measurement of the individual’s subjective expertise was taken from the work of Flynn and Goldsmith (1999) and was adapted to the French context by Lombart (2004) (α = 0.924).⁹ Finally, measurement of the perceived quality of the product is taken from the work of Morales and Fitzsimmons (2007), who use an ad hoc item.

Findings

We verify our manipulation by carrying out an ANOVA with the manipulation item check as a dependent variable and the condition of similarity to a new product as an independent variable. The result of the test is significant: similarity to the new product is higher under the condition of strong similarity (M = 4.35) than that of weak similarity (M = 3.35; F (1.128) = 12.694; p = 0.00). However, the product is liked as much under the condition of strong similarity (M = 4.05) as under the condition of weak similarity (M = 4.22; p = 0.56). The mean values, correlations and distribution parameters of the quantitative variables are shown in Appendix 6.

We test hypotheses H4a and H4b using model 1 of the PROCESS macro. The results obtained show that perceived physical risk is lower when the product looks new (b = −1.71; t = −2.50; p = 0.01). However, this result is affected by the individual’s level of subjective expertise: the interaction between similarity and expertise is significant (b = 0.49; t = 2.31; p = 0.02). Since subjective expertise is a variable without a focal value – the scale is non-significant and does not include any specific values associated with change in behaviour or the consumer’s perception – following the recommendations of Cadario and Parguel (2014), we explore this interaction with a floodlight analysis (Spiller et al., 2013). The Johnson–Neyman point shows that the similarity effect only exists if the individual has little expertise, with an average level below 2.0953. Beyond that, the effect becomes insignificant (Appendix 7). Hypothesis H4a is therefore confirmed. Testing hypothesis H4b also reveals significant interaction between similarity and subjective expertise (b = –0.51; t = –3.02; p = 0.00). However, contrary to our expectations, the Johnson–Neyman point reveals that the effect of similarity occurs only if the individual is an expert, with an average level of expertise higher than 2.7773 (Appendix 7). For these individuals, similarity results in a sharp drop in perceived quality. Despite this interesting finding, H4b has to be rejected.

We test hypotheses H5a and H5b using model 9 of the PROCESS macro, which differs from model 7 in the action of two moderators on the link between the independent and mediating variables (subjective expertise and sensitivity to interpersonal disgust). Using the same methodological approach to test for mediating effects as in Study 1, we find that the indirect effect of similarity on purchase intention through perceived physical risk (a × b = 0.01) is not significant. The associated 95% confidence interval includes 0 [−0.02, 0.14] regardless of the levels of expertise in the product category and sensitivity to interpersonal disgust. The moderated mediation indexes are not significant (I = −0.01 and I = 0.02, respectively), with confidence intervals that include 0 ([−0.11, 0.03] and [−0.02, 0.06], respectively). Hypothesis H5a is therefore disconfirmed. However, there is a significant mediating effect of similarity on purchase intention via perceived quality (a × b = −0.22), with a 95% confidence interval that does not include 0 [−0.49, −0.02]. In accordance with the results of testing hypothesis H4b, this mediation occurs only among individuals with greater expertise and for whom similarity, by decreasing perceived quality, results in a lower purchase intention (I = −0.25 with a confidence interval that excludes 0 [−0.47, −0.07]. Hypothesis H5b is confirmed.

Discussion

The confirmation of hypothesis H4a answers our third research question by showing that the similarity heuristic applies to non-expert individuals, for whom greater similarity reduces perceived physical risk. This finding reveals, for the first time in an online sales context, the utility of the similarity law for limiting the effects of negative physical contagion. Nevertheless, whereas Rozin and Nemeroff (2002) show that the effects of the law of similarity influence behaviour, our second experiment does not reveal any mediating effect of similarity on purchase intention through perceived physical risk. A possible explanation for this could be that the product concerned (a headphone) has a lower level of bodily intimacy. Thus perceived physical risk probably plays a lesser role.

Hypothesis H4b, which has not been confirmed, points to a particularly interesting counterintuitive effect: similarity does not increase perceived quality for non-experts, and decreases it for those with the most expertise. Non-experts would therefore not use packaging as a heuristic to evaluate the quality of the product. On the other hand, for experts relying on multi-criteria evaluation, similarity seems to produce an inconsistency effect between the ad for a second-hand product and resemblance to a new product. We will return to this point in the conclusion. H5b is confirmed, with perceived quality mediating the effect of similarity on purchase intention for expert individuals, thus highlighting the important role played by perceived quality in the formation of purchase intention for a used electronic product (Ghose, 2009).

Finally, the disconfirmation of H5b shows that individuals who are very sensitive to interpersonal disgust and who are not experts do not seem to perceive a greater physical risk if the second-hand product is presented without packaging. The effect of negative physical contagion would therefore not be stronger or weaker according to the individual’s sensitivity to interpersonal disgust, a finding that once again suggests that this effect has a cognitive mechanism.

A summary of the results is provided in Appendix 8.

Conclusion

Theoretical contributions

Our research brings to light a negative physical contagion effect in the context of second-hand shopping online. Thus, even without direct contact with the product and simply being shown a photo of it, the fact of being aware that the previous owner has been in contact with the product increases perceived physical risk; and in turn perceived risk influences purchase intention. Our research thus contributes to the literature on negative physical contagion, which hitherto has been limited to offline purchase, by showing that such contagion is also possible online. The observed effect allows us to identify a new antecedent of negative physical contagion, namely a photo showing contact between the product and an unknown person.

Study 1 showed the central role played by perceived physical risk in purchase intent, as a mediating variable between contact salience and purchase intention. This result is consistent with the literature on both online shopping and on second-hand shopping. On the other hand, this finding differs from the marketing literature with regard to product contagion. Whereas Argo et al. (2006) find that contact salience has a direct effect on purchase intent, our results suggest that it is only indirect. In addition, contrary to the existing literature that emphasizes the role of disgust in the physical contagion effect (Argo et al., 2006; Di Muro and Noseworthy, 2013; Roux, 2004a), our results indicate that disgust would not play a role in the formation of perceived physical risk, since the moderation effect through sensitivity to interpersonal disgust was not evident in either Study 1 or Study 2. Although an emotional effect through feelings of disgust is not totally excluded, our findings argue for a cognitive mechanism for the effect of negative physical contagion in the context of second-hand purchase online. These results are consistent with the literature in psychology, which identifies two possible mechanisms to explain the effects of negative physical contagion. However, these may stem from an emotional mechanism linked to disgust generated by the object concerned (McKay and Tsao, 2005; Moretz and Mckay, 2008; Olatunji et al., 2004). However, they may be based on a cognitive mechanism related to beliefs about and assessment of risks arising from the second-hand product (Rachman, 2006). Adams et al. (2013) show that both mechanisms are involved in the perception of negative physical contagion, but that one or other of them is often predominant. According to Cisler et al. (2011) and Adams et al. (2013), the predominance of a mechanism depends on the distance between the individual and the contagious object: the mechanism is more emotional if the contact is direct, and more cognitive if the contact is indirect. In a context of online shopping, contact with the item is mediated by the screen of the device employed, and the perceived distance is therefore probably greater.

Moreover, our research contributes to a better understanding of the functioning of the law of similarity, in the presence of contact salience induced by the nature of the product. There are currently few studies on the law of similarity. Such studies as exist are confined to the psychology literature (Nemeroff and Rozin, 1994; Rozin et al., 1986) and deal mainly with contagion through similarity: an object that looks like another object that is contagious can give rise to contagion. We, here, show the possibility of a reduction of the negative effect of physical contagion on perceived risk through similarity with a new product (Rozin, 1994; Rozin and Nemeroff, 2002). However, this effect only occurs for certain individuals who are not expert in the product category. We thus confirm the importance of the individual variable of perceived expertise in the product category, which influences different forms of reasoning regarding the assessment of risk and perceived quality, and therefore affects the formation of purchase intention.

For individuals who are expert in the product category, the effect of similarity is unexpected, since it tends to reduce the perceived quality of the product and thus the purchase intention. This counterintuitive result can be explained by how expert consumers reason. Because they emphasize the cognitive side more (Alba and Hutchinson, 1987), they over-process information that is vividly presented with a view to attracting attention (Kim et al., 1991). This is the paradox of the experts: even though they are able to process more information, and more complex information than novices, they may also overestimate the importance of certain information, and over-interpret it. The inconsistency between the fact that the product is second-hand and that it is presented in packaging similar to that of new products can then generate distrust: experts may perceive presentation of this kind as an attempt to mask the real state of the product. This mistrust on the part of experts towards certain attributes of the product has already been underlined in other studies, particularly those concerning environmental communication. For example, the use of the colour green can harm the ecological image of the brand for expert consumers, who perceive manipulative intent on the part of the advertiser (Benoit-Moreau et al., 2010).

Managerial contributions

From a managerial standpoint, this research highlights the role of negative physical contagion in buying second-hand goods online. A number of practical recommendations can thus be made for professionals or those who market second-hand products on the Internet.

It is important to reduce the salience of contact, by avoiding making it explicit (e.g. not using a photo of the product being worn or not mentioning contact with the product in the description). While this practice seems to be widely used on professional websites selling second-hand products, it may nonetheless conflict with the strategies of certain websites that aim to establish a personal link between sellers and buyers (example from Vestiairecollective). Moreover, on sites selling products by putting individuals in touch with each other, like Leboncoin in France, or eBay, the product is sometimes shown in contact with the owner. Our research findings strongly suggest that sellers should avoid these practices.

However, similarity of the second-hand product to a new product can reduce the effect of the negative physical contagion on perceived risk for non-expert individuals. This finding validates strategies of presenting second-hand products as new (use of packaging, presence of labels, presence of generic visuals identical to those of new products, etc.). An ethical question may, however, arise if these methods of presentation lead customers to forget that the product is in fact second-hand. But the results of our pre-tests seem to show that this confusion does not occur, with respondents remaining well aware that the product is not new.

In this research we also emphasize the fact that the similarity of the used product to the new one can have a negative effect on perceived quality for the most expert customers. Consequently, websites using this presentation strategy should do so with discernment, depending on the type of product and the consumers targeted. For complex technological products that target experts, assessment of physical risk and quality will not be based on the similarity between the second-hand product and a new one. It is therefore unnecessary, or potentially counterproductive, to use these presentation strategies. However, for mid-level product categories, some sites target the general public and this presentation strategy could limit the perception of physical risk. This is all the more relevant for products such as shoes and clothing with a high degree of physical intimacy. Though this strategy is already frequently used on professional websites, those linking individuals could encourage sellers to follow it. It is important, however, for obvious ethical reasons, to state clearly and visibly that the product is second-hand.

In view of these results and these ethical considerations, a further recommendation would be to act directly on the reduction of perceived risk, without emphasizing similarity, by explicitly mentioning that the products have been cleaned, thus underlining the non-harmful nature of potential residues, or using the testimonials of consumers who have been happy with products having high degree of intimacy, even when second-hand.

Limitations and perspectives

This research has a number of limitations, that in turn open up avenues of research. First of all, we chose here to study only one dependent variable, namely purchase intention. This decision was linked to a desire to investigate the mediating and moderating effects in greater depth, and to be able to compare the results with the existing literature. However, the inclusion of other dependent variables in the model could have allowed more direct effects to be observed. Similarly, we limited ourselves to a small number of control variables.

The stimuli used also have their limitations. The mean purchase intention scores for the products we chose are relatively low. We also used a dummy website to control for sources of variance exogenous to the model being tested. However, doing so may have created suspicion towards an unknown site. In future studies, other products as well as genuine websites will need to be used to establish the external validity of our model.

Furthermore, we decided to work with a young cohort and to control for the age of participants. According to Dehling (2013), younger consumers are more likely to buy second-hand products with high levels of bodily intimacy. Hence, our research does not allow us to study a potential age or generational effect: it is possible that the lack of any direct effect of contact salience on the intention to purchase online a product that has been handled/worn is linked to the age of the participants. It would therefore be particularly worthwhile carrying out a study with older participants, who could potentially be more susceptible to contagion effects.

The lack of a direct link between contact salience and purchase intention also suggests new lines of research. Indeed, the lack of a direct link could be accounted for by a competing mediator with effects that are opposed to those of perceived physical risk. The literature leads us, for example, to consider the role of embodied mental simulation: contact salience could allow consumers to better imagine themselves using the product and thus increase their purchase intention (Elder and Krishna, 2012; Kim, 2017). The simultaneous examination of perceived physical risk and embodied simulation would thus constitute an interesting extension of our research.

Finally, the current growth of so-called collaborative consumption, including the purchase of second-hand products, as well as products rented or exchanged on the Internet, calls for the development of research on all products that are shared and therefore handled by others, via online platforms. It would be particularly interesting to see whether the identification of the negative physical contagion effect can be replicated in these contexts.

Footnotes

Appendix 1

Main features of buying second-hand products online and offline.

		Offline second-hand purchase	Online second-hand purchase
Characteristics of distribution channels	Access to channels and extent of commercial offering	Access to channels is restricted by their location and opening times; limited offering (Singh et al., 2014).	As for online shopping in general, 24-hour purchase from home (Wolfinbarger and Gilly, 2001). Access to a wider product offering (Cameron and Galloway, 2005; Ghose et al., 2006; Singh et al., 2014).
	Boundary between new and second-hand products	Second-hand channels are often specific and distinct from new channels (Guiot and Roux, 2010).	The boundaries between new and used products may be blurred: presentation of these two types of products is similar, consisting of photos and descriptions (Gregg and Walczak, 2008). In addition, many sites offer both new and used – for example, eBay or Amazon (Kauffman and Wood, 2006; Walczak et al., 2006; Ertz, 2015).
	Information available	Information collected in physical distribution channels may be difficult to obtain (Helme-Guizon, 2001).	Abundant information: second-hand product offerings are more standardized and at least contain information included in the photos and the description of the object (Gregg and Walczak, 2008). Further information may also be available on the seller’s history (sales number, reputation index, etc.) (Li et al., 2009; Ghose, 2009).
Consumption experience at sales outlets	Sensory stimulation	Perception of the product involves several senses: the product can be seen, touched and smelled (Solomon, 2008).	Perception of the product is mediated by a screen (Helme-Guizon, 2001): sensory stimulation is therefore only visual, limited to descriptions and photographs that do not allow the product to be seen from all angles (Abdul-Ghani et al., 2011).
Consumption experience at sales outlets	Social interaction	Interaction is limited to a particular geographical area and specific social circles (Belk et al., 1988). The identity of the seller is obvious, the exchange is based on verbal and nonverbal communication (Sherry, 1990; Abdul-Ghani et al., 2011).	The Internet promotes interaction between strangers and provides access to other social circles and other national or international geographic areas (Botsman and Rogers, 2010). Sellers and buyers have no face-to-face contact (Rafaeli and Noy, 2002; Yen and Lu, 2008), and are not physically present at the same place and at the same time (Chakravarti et al., 2002; Cui et al., 2008).
Evaluation of the commercial offering	Perceived physical risk	In physical channels, second-hand products can be inspected or even tested before purchase (Cameron and Galloway, 2005), which reduces perceived performance and financial risk (Roux, 2004a).	Risk is exacerbated in online shopping (Tan, 1999), particularly in terms of financial and performance risk (Biswas and Biswas, 2004). Buying second-hand online does not allow the product to be directly inspected (Lai et al., 2008) and the photos provided are not always sufficient to estimate the condition of the item (Ertz, 2015).
Evaluation of the commercial offering	Cognitive effort	Less textual information, contact with the seller is possible (Alba et al., 1997; Helme-Guizon, 2001).	Consumers using C2C online channels seek more information to compensate for perceived risk (Anderson and Zahaf, 2007). The abundance of information complicates decision-making and requires cognitive effort (Ariely and Simonson, 2003), and may even lead to information overload (Jacoby and Kaplan, 1972). In the case of online marketplaces, the decision process tends to be rational, based in particular on an assessment of the quality of the product and its price (Wang et al., 2002).

Appendix 2

Appendix 3

Measurement instruments.

Variable	Reference	Assertions	Format	Reliability index
Variable	Reference	Assertions	Format	Study 1	Study 2
Purchase intention	Argo et al. (2006)	Assuming you need them, how likely are you to buy these running shoes?	Very unlikely – Very likely (1– 7)	NA	NA
Perceived physical risk	Stone and Grönhaug (1993)	One of my concerns about buying this product is hygiene.	Likert 1–7	α = 0.703	α = 0.886
		I wonder whether this product will have harmful physical effects such as the transmission of germs or diseases.	Likert 1–7
		I am worried about the bodily dangers associated with using this product.	Likert 1–7
Perceived performance risk	Stone and Grönhaug (1993)	I’m worried about whether these running shoes will really be in good condition.	Likert 1–7	α = 0.584	α = 0.701
		I wonder how much I can really rely on these running shoes.	Likert 1–7
		I’d be worried that these running shoes do not have the qualities I’d expect.	Likert 1–7
Perceived financial risk	Stone and Grönhaug (1993)	If I bought this product, I’d be worried about making a bad investment.	Likert 1–7	α = 0.627	α = 0.868
		If I bought this product, I’d be worried about not getting my money’s worth.	Likert 1–7
		Purchasing this product could be a waste of money.	Likert 1–7
Sensitivity to interpersonal disgust	Haidt et al. (1994)	I’d find it disgusting to drink, even inadvertently, from a glass that a stranger had just drunk before me.	Likert 1–7	α = 0.773	α = 0.767
		In general, I avoid letting any part of my body touch the toilet seat in public toilets.	Likert 1–7
		I would not go to my favourite restaurant if I found out that the cook had a cold.	Likert 1–7
		I would not agree to someone I don’t like sleeping in my bed.	Likert 1–7
		I cannot stand the idea of wearing clothes, even if washed, that have belonged to someone else.	Likert 1–7
		I could not put a bank note that people have previously handled between my lips	Likert 1–7
Subjective expertise	Lombart (2004)	I know a lot about headphones.	Likert 1–7		α = 0.924
		In my circle of friends, I am an expert on headphones.	Likert 1–7
		I don’t feel very well informed about headphones.	Likert 1–7
Perceived quality	Morales and Fitzsimmons (2007)	The quality of this product is good.	Likert 1–7		NA
Sustained involvement	Strazzieri (1994)	What do you think about buying running shoes in general?	Not important – important (1–7)	α = 0.897	α = 0.809
			Not interested – interested (1–7)
			Not attracted – attracted (1–7)

Appendix 4

Appendix 5

Results of Study 1.

No.	Hypothesis	Test	Result	Summary
H1a	Contact salience positively influences perceived physical risk	General linear model	p = 0.049	C
H1b	Contact salience positively influences perceived performance risk	General linear model	n.s.	D
H1c	Contact salience positively influences perceived financial risk	General linear model	n.s.	D
H2a	The influence of contact salience on purchase intention is mediated by perceived physical risk	PROCESS macro model 7	(a × b = −0.16) with confidence interval without 0 [−0.56, −0.01]	C
H2b	The influence of contact salience on purchase intention is mediated by perceived performance risk	PROCESS macro model 7	(a × b = 0.07) n.s.	D
H2c	The influence of contact salience on purchase intention is mediated by perceived financial risk	PROCESS macro model 7	(a × b = 0.01) n.s.	D
H3	Sensitivity to interpersonal disgust moderates the mediation of the influence of contact salience on purchase intention by perceived physical risk	PROCESS macro model 7	Moderated mediation index (I = −0.03) n.s.	D

C: confirmed, D: disconfirmed.

Appendix 6

Appendix 7

Appendix 8

Results of Study 2.

No.	Hypothesis	Test	Result	Summary
H4a	The effect of similarity on perceived physical risk is moderated by the level of expertise: similarity negatively influences perceived physical risk for non-experts; this effect is not observed for experts	PROCESS macro model 1	Significant interaction (p = 0.02). Significant effect for expertise lower than 2.0953, n.s. for expertise higher than this threshold.	C
H4b	The effect of similarity on perceived quality is moderated by the level of expertise: similarity positively influences perceived quality for non-experts; this effect is not observed for experts pour les experts	PROCESS macro model 1	Significant interaction (p = 0.00). Significant effect for expertise higher than 2.7773, n.s. for expertise lower than this threshold.	D
H5a	Perceived physical risk mediates the effect of similarity to a new product on purchase intention; this mediation is moderated by the perceived expertise and sensitivity to interpersonal disgust	PROCESS macro model 9	(a × b = 0.01) n.s. Moderated mediation indices (I = −0.01; I = 0.02) n.s.	D
H5b	Perceived quality mediates the effect of the similarity to a new product on purchase intention; this mediation is moderated by perceived expertise	PROCESS macro model 9	(a × b = −0.22) with confidence interval without 0 [−0.49, −0.02]. Moderated mediation index (I = −0.25) with confidence interval without 0 [−0.47, −0.07]	C

C: confirmed; D: disconfirmed.

Notes

References

Abbey

Meloy

Guide

VDR

Atalay

(2015) Remanufactured products in closed loop supply chains for consumer goods. Production and Operations Management 24(3): 488–503.

Abdul-Ghani

Hyde

Marshall

(2011) Emic and etic interpretations of engagement with a consumer-to-consumer online auction site. Journal of Business Research 64(10): 1060–1066.

Ackerman

(2017) Assuring me that it is as ‘Good as New’ just makes me think about how someone else used it: Examining consumer reaction toward marketer-provided information about second-hand goods. Journal of Consumer Behaviour 16(3): 233–241.

Adams

Cisler

Brady

Lohr

Olatunji

(2013) Preliminary psychometric evidence for distinct affective and cognitive mechanisms mediating contamination aversion. Journal of Psychopathology and Behavioral Assessment 35(3): 375–388.

Agarwal

Teas

(2001) Perceived value: Mediating role of perceived risk. Journal of Marketing Theory and Practice 9(4): 1–14.

Akerlof

(1970) The market for ‘lemons’: Quality uncertainty and the market mechanism. Quarterly Journal of Economics 84(3): 488–500.

Alba

Hutchinson

(1987) Dimensions of consumer expertise. Journal of Consumer Research 13(4): 411–454.

Alba

Lynch

Weitz

Janiszewski

Lutz

Sawyer

Wood

(1997) Interactive home-shopping: Consumer, retailer, and manufacturer incentives to participate in electronic marketplaces. Journal of Marketing 61(3): 38–53.

Anderson

Zahaf

(2007) Profiling consumer-to-consumer and business-to-consumer buyers: Who buys what? International Journal of Business Research 7(6): 85–99.

10.

Argo

Dahl

Morales

(2006) Consumer contamination: How consumers react to products touched by others. Journal of Marketing 70(2): 81–94.

11.

Ariely

Simonson

(2003) Buying, bidding, playing, or competing? Value assessment and decision dynamics in online auctions. Journal of Consumer Psychology 13(1): 113–123.

12.

Ayadi

(2010) L’expérience de prise de risque par les consommateurs: rôle des émotions anticipées positives. Recherche et Applications en Marketing 25(2): 31–49.

13.

Bardhi

Eckhardt

(2012) Access-based consumption: The case of car sharing. Journal of Consumer Research 39(4): 881–898.

14.

Baron

Kenny

(1986) The moderator-mediator variable distinctive in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology 51(6): 1173–1182.

15.

Bauer

(1960) Consumer behavior as risk taking. In: Hancock

(ed.) Dynamic Marketing for a Changing World. Chicago, IL: American Marketing Association, pp. 389–398.

16.

Belk

Sherry

Wallendorf

(1988) A naturalistic inquiry into buyer and seller behavior at a swap meet. Journal of Consumer Research 14(4): 449–470.

17.

Benoit-Moreau

Larceneux

Parguel

(2010) La communication sociétale: entre opportunités et risques d’opportunisme. Décisions Marketing 59: 75–78.

18.

Bettman

Park

(1980) Effects of prior knowledge and experience and phase of choice process on consumer decision processes: A protocol analysis. Journal of Consumer Research 7(4): 234–248.

19.

Bezançon

(2012) Pourquoi les consommateurs n’achètent-ils pas d’occasion? Une analyse exploratoire. In: Proceedings of the 28th international congress of the French Marketing Association, Brest, 5 May.

20.

Biswas

(2004) The diagnostic role of signals in the context of perceived risks in online shopping: Do signals matter more on the Web? Journal of Interactive Marketing 18(3): 30–45.

21.

Botsman

Rogers

(2010) What’s Mine Is Yours: The Rise of Collaborative Consumption. New York: Harper Collins.

22.

Cadario

Parguel

(2014) Reconsidérer la discrétisation des variables quantitatives: vers une nouvelle analyse de modération dans la recherche expérimentale. Recherche et Applications en Marketing 29(4): 1–18.

23.

Cameron

Galloway

(2005) Consumer motivations and concerns in online auctions: An exploratory study. International Journal of Consumer Studies 29(3): 181–192.

24.

Castro

Morales

Nowlis

(2013) The influence of disorganized shelf displays and limited product quantity on consumer purchase. Journal of Marketing 77(4): 118–133.

25.

Chakravarti

Greenleaf

Sinha

Cheema

Cox

Friedman

Isaac

Mitchell

Rapoport

Rothkopf

Srivastava

Zwick

(2002) Auctions: Research opportunities in marketing. Marketing Letters 13(3): 281–296.

26.

Chaudhuri

(1997) Consumption emotion and perceived risk: A macro-analytic approach. Journal of Business Research 39(2): 81–92.

27.

Chaudhuri

(1998) Product class effects on perceived risk: The role of emotion. International Journal of Research in Marketing 15(2): 157–168.

28.

Chaudhuri

(2002) A study of emotion and reason in products and services. Journal of Consumer Behaviour 1(3): 267–279.

29.

Chen

Dubinsky

(2003) A conceptual model of perceived customer value in e-commerce: A preliminary investigation. Psychology & Marketing 20(4): 323–347.

30.

Cisler

Adams

Brady

Bridges

Lohr

Olatunji

(2011) Unique affective and cognitive processes in contamination appraisals: Implications for contamination fear. Journal of Anxiety Disorders 25(1): 28–35.

31.

Cui

Lai

Liu

CKW

(2008) Research on consumer behaviour in online auctions: Insights from a critical literature review. Electronic Markets 18(4): 345–361.

32.

Dehling

(2013) La mise en soi de l’objet de l’Autre, des concepts de possessions et d’appropriation dans le contexte de la consommation d’occasion. PhD Thesis, École des Hautes Etudes en Sciences Sociales, Paris.

33.

Di Muro

Noseworthy

(2013) Money isn’t everything but it helps if it doesn’t look used: How the physical appearance of money influences spending. Journal of Consumer Research 39(6): 1330–1342.

34.

Doleac

Stein

(2012) The visible hand: Race and online market outcomes. Economic Journal 123(572): 469–492.

35.

Donthu

Garcia

(1999) The Internet shopper. Journal of Advertising Research 39(3): 52–58.

36.

Elder

Krishna

(2012) The ‘visual depiction effect’ in advertising: Facilitating embodied mental simulation through product orientation. Journal of Consumer Research 38(2): 988–1003.

37.

Ertz

(2015) Du Web 2.0 à la seconde vie des objets. Le rôle de la technologie Internet. San Francisco, CA: Grin Publishing.

38.

Flynn

Goldsmith

(1999) A short, reliable measure of subjective knowledge. Journal of Business Research 46(1): 57–66.

39.

Forsythe

Shi

(2003) Consumer patronage and risk perceptions in Internet shopping. Journal of Business Research 56(11): 867–875.

40.

Frazer

(1959 [1890]) The New Golden Bough: A Study in Magic and Religion. New York: Macmillan.

41.

Gabbott

(1991) The role of product cues in assessing risk in second-hand markets. European Journal of Marketing 25(9): 38–50.

42.

Gerstner

(1985) Do higher prices signal higher quality? Journal of Marketing Research 22(2): 209–215.

43.

Ghose

(2009) Internet exchanges for used goods: An empirical analysis of trade patterns and adverse selection. MIS Quarterly 33(2): 263–291.

44.

Ghose

Smith

Telang

(2006) Internet exchanges for used books: An empirical analysis of product cannibalization and welfare impact. Information Systems Research 17(1): 3–19.

45.

Gil

Rousset

Droit-Volet

(2009) How liked and disliked foods affect time perception. Emotion 9(4): 457–463.

46.

Goukens

Dewitte

Pandalaere

Warlop

(2007) Wanting a bit(e) of everything: Extending the valuation effect to variety seeking. Journal of Consumer Research 34(3): 386–394.

47.

Gregg

Walczak

(2008) Dressing your online auction business for success: An experiment comparing two e-bay businesses. MIS Quarterly 32(3): 653–670.

48.

Gregson

Crewe

(2003) Second-Hand Cultures. Oxford: Berg.

49.

Grewal

Gotlieb

Marmorstein

(1998) The moderating effects of message framing and source credibility on the price-perceived risk relationship. Journal of Consumer Research 21(1): 145–153.

50.

Guiot

Roux

(2010) A second-hand shoppers’ motivation scale: Antecedents, consequences and implications for retailers. Journal of Retailing 86(4): 383–399.

51.

Haidt

McCauley

Rozin

(1994) Individual differences in sensitivity to disgust: A scale sampling seven domains of disgust elicitors. Personality and Individual Differences 16(5): 701–713.

52.

Haidt

McCauley

Rozin

(2002) The Disgust Scale, version 2. Retrieved from: http://www.people.virginia.edu/-jdh6n/disgustscale.html

53.

Hair

Black

Babin

Anderson

(2010) Multivariate Data Analysis. London: Pearson.

54.

Hayes

(2013) Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach, 1st edition. New York: The Guilford Press.

55.

Hayes

(2018) Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach, 2nd edition. New York: The Guilford Press.

56.

Helme-Guizon

(2001) Le comportement du consommateur sur un site marchand est-il fondamentalement différent de son comportement en magasin? Proposition d’un cadre d’appréhension de ses spécificités. Recherche et Applications en Marketing 16(3): 25–38.

57.

Hodson

Costello

(2007) Interpersonal disgust, ideological orientations, and dehumanization as predictors of intergroup attitudes. Psychological Science 18(8): 691–698.

58.

Huang

Ackerman

Newman

(2017) Catching (up with) magical contagion: A review of contagion effects in consumer contexts. Journal of the Association for Consumer Research 2(4): 430–443.

59.

Jacoby

Kaplan

(1972) The components of perceived risk. Advances in Consumer Research 12: 287–291.

60.

Kalwani

Silk

(1982) On the reliability of predictive validity of purchase intention measures. Marketing Science 1(3): 243–286.

61.

Kapitan

Bhargave

(2013) Navigating residue sensitivity in the used goods marketplace. Psychology & Marketing 30(4): 305–317.

62.

Kaplan

Szybillo

Jacoby

(1974) Components of perceived risk in product purchase: A cross validation. Journal of Applied Psychology 59(3): 287–291.

63.

Kauffman

Wood

(2006) Doing their bidding: An empirical examination of factors that affect a buyer’s utility in Internet auctions. Information Technology and Management 7(3): 171–190.

64.

Keh

Pang

(2010) Customer reactions to service separation. Journal of Marketing 74(2): 55–70.

65.

Kim

(2017) The ownership distance effect: The impact of traces left by previous owners on the evaluation of used goods. Marketing Letters 28(4): 591–605.

66.

Kim

Kardes

Herr

(1991) Consumer expertise and the vividness effect: Implications for judgment and inference. In: Holman

Solomon

(eds) NA – Advances in Consumer Research, vol. 18. Provo, UT: Association for Consumer Research, pp. 90–93.

67.

Kwon

Lennon

(2009) What induces online loyalty? Online versus offline brand images. Journal of Business Research 62(5): 557–564.

68.

Lai

Win

(2008) A qualitative approach for conceptualizing consumer decision-making in online auctions. In: Lee

Soman

(eds) NA – Advances in Consumer Research, vol. 35. Duluth, MN: Association for Consumer Research, pp. 319–324.

69.

Srinivasan

Sun

(2009) Internet auction features as quality signals. Journal of Marketing 73(1): 75–92.

70.

Lombart

(2004) Le butinage: proposition d’une échelle de mesure. Recherche et Applications en Marketing 19(2): 1–30.

71.

McKay

Tsao

(2005) A treatment most foul: Handling disgust in cognitive-behavior therapy. Journal of Cognitive Psychotherapy 19(4): 355–367.

72.

Maheswaran

(1994) Country of origin as a stereotype: Effects of consumer expertise and attribute strength on product evaluations. Journal of Consumer Research 21(2): 354–365.

73.

Maheswaran

Sternthal

(1990) The effects of knowledge, motivation, and type of message on ad processing and product judgments. Journal of Consumer Research 17(1): 66–73.

74.

Maheswaran

Mackie

Chaiken

(1992) Brand name as a heuristic cue: The effects of task importance and expectancy confirmation on consumer judgments. Journal of Consumer Psychology 1(4): 317–336.

75.

Mauss

(1950 [1902]) Théorie générale de la magie in sociologie et anthropologie. Paris: PUF.

76.

Merle

(2007) La valeur perçue de la customisation de masse: proposition et test d’un modèle conceptuel intégrateur. PhD Thesis, CERGAM, Paul-Cézanne University (Aix-Marseille III, IAE), Aix-en-Provence.

77.

Morales

Fitzsimmons

(2007) Product contagion: Changing consumer evaluations through physical contact with ‘disgusting’ products. Journal of Marketing Research 44(2): 272–283.

78.

Moretz

Mckay

(2008) Disgust sensitivity as a predictor of obsessive-compulsive contamination symptoms and associated cognitions. Journal of Anxiety Disorders 22(4): 707–715.

79.

Nemeroff

Rozin

(1994) The contagion concept in adult thinking in the United States: Transmission of germs and of interpersonal influence. Ethos: Journal of the Society for Psychological Anthropology 22(2): 158–186.

80.

Olatunji

Sawchuk

Lohr

de Jong

(2004) Disgust domains in the prediction of contamination fear. Behaviour Research and Therapy 42(1): 93–104.

81.

Olatunji

Williams

Tolin

Abramowitz

Sawchuk

Lohr

Elwood

. (2007) The disgust scale: Item analysis, factor structure, and suggestions for refinement. Psychological Assessment 19(3): 281–297.

82.

O’Reilly

Rucker

Hughes

Gorang

Hand

(1984) The relationship of psychological and situational variables to usage of a second-order marketing system. Journal of the Academy of Marketing Science 12(3): 53–76.

83.

Orth

Campana

Malkewitz

(2014) Formation of consumer price expectation based on package design: Attractive and quality routes. Journal of Marketing Theory and Practice 18(1): 23–40.

84.

Pantoja

Bakpayev

Rossi

Yoon

(2016) Old, but gold! The role of aging stereotypes on consumer’s purchase intentions in second-hand marketplaces. In: Moreau

Puntoni

(eds) NA – Advances in Consumer Research, vol. 44. Duluth, MN: Association for Consumer Research, pp. 580–581.

85.

Preacher

Hayes

(2008) Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods 40(3): 879–891.

86.

Rachman

(2006) Fear of Contamination: Assessment and Treatment. New York: Oxford University Press Inc.

87.

Rafaeli

Noy

(2002) Online auctions, messaging, communication and social facilitation: A simulation and experimental evidence. European Journal of Information System 11(3): 196–207.

88.

Raju

Lonial

Mangold

(1995) Differential effects of subjective knowledge, objective knowledge, and usage experience on decision making: An exploratory investigation. Journal of Consumer Psychology 4(2): 153–180.

89.

Rao

Monroe

(1989) The effect of price, brand name, and store name on buyers’ perceptions of product quality: An integrative review. Journal of Marketing Research 36(3): 351–357.

90.

Reynolds-McIlnay

Morrin

Nordfält

(2017) How product-environment brightness contrast and product disarray impact consumer choice in retail environments. Journal of Retailing 93(3): 266–282.

91.

Rieunier

(2000) L’influence de la musique d’ambiance sur le comportement du consommateur sur le lieu de vente. PhD Thesis, DMSP, Paris IX-Dauphine University, Paris.

92.

Roselius

(1971) Consumer rankings of risk reduction methods. Journal of Marketing 35(1): 56–61.

93.

Roux

(2004a) Buying second-hand clothes: An exploratory approach using differences in consumers’ disgust sensitivity. In: Proceedings of the 33rd congress of EMAC, Murcia.

94.

Roux

(2004b) Les orientations envers l’achat d’occasion, le rôle du risque perçu et de l’expertise: Quelles implications pour la distribution? Décisions Marketing 35: 9–21.

95.

Roux

Guiot

(2008) Une mesure des motivations envers l’achat d’occasion, leurs antecedents et leurs consequences. Recherche et Applications en Marketing 23(4): 63–95.

96.

Roux

Korchia

(2006) Am I what I wear? An exploratory study of symbolic meanings associated with second-hand clothing. Advances in Consumer Research 33(1): 29–35.

97.

Rozin

(1994) Manger magique. In: Fischler

(eds) Manger magique. Aliments sorciers, croyances comestibles (Autrement, Mutations/Mangeurs).

98.

Rozin

Nemeroff

(2002) Sympathetic magical thinking: The contagion and similarity heuristics. In: Gilovich

Griffin

Kahneman

(eds) Heuristics and Biases: The Psychology of Intuitive Judgment. Cambridge: Cambridge University Press, pp. 201–216.

99.

Rozin

Haidt

McCauley

(2000) Disgust. In: Lewis

Haviland-Jones

(Coords) Handbook of Emotions, 2nd edition. New York: The Guilford Press, pp. 637–653.

100.

Rozin

Millman

Nemeroff

(1986) Operation of the laws of sympathetic magic in disgust and other domains. Journal of Personality and Social Psychology 50(4): 703–712.

101.

Sherry

(1990) A sociocultural analysis of a Midwestern American flea market. Journal of Consumer Research 17(1): 113–130.

102.

Singh

Ratchford

Prasad

(2014) Offline and online search in used durables markets. Journal of Retailing 90(3): 301–320.

103.

Solomon

(2008) Consumer Behavior: Buying, Having and Being, 8th edition. New York: Prentice Hall.

104.

Spiller

Fitzsimons

Lynch

Mcclelland

(2013) Spotlights, floodlights, and the magic number zero: Simple effects tests in moderated regression. Journal of Marketing Research 50(2): 277–288.

105.

Srivastava

Mitra

(1998) Warranty as a signal of quality: The moderating effect of consumer knowledge on quality evaluations. Marketing Letters 9(4): 327–336.

106.

Stone

Grönhaug

(1993) Perceived risk: Further considerations for the marketing discipline. European Journal of Marketing 27(3): 39–50.

107.

Strazzieri

(1994) Mesurer l’implication durable vis-à-vis d’un produit indépendamment du risque perçu. Recherche et Applications en Marketing 9(1): 73–91.

108.

Suwelack

Hogreve

Hoyer

(2011) Understanding money-back guarantees: Cognitive, affective, and behavioral outcomes. Journal of Retailing 87(4): 462–478.

109.

Tan

(1999) Strategies for reducing consumers’ risk aversion in Internet shopping. Journal of Consumer Marketing 16(2): 163–180.

110.

Tylor

(1974 [1871]) Primitive Culture: Researches into the Development of Mythology, Philosophy, Religion, Art and Custom. New York: Gordon Press.

111.

Van der Heijden

Verhagen

(2004) Online store image: Conceptual foundations and empirical measurement. Information & Management 41(5): 609–617.

112.

Volle

(1995) Le concept de risque perçu en psychologie du consommateur: antécédents et statut théorique. Recherche et Applications en Marketing 10(1): 39–56.

113.

Walczak

Gregg

Berrenberg

(2006) Market decision making for online auction sellers: Profit maximization or socialization. Journal of Electronic Commerce Research 7(4): 199–220.

114.

Wang

ETG

Tai

(2002) A study of online auction sites in Taiwan: Product, auction rule, and trading type. International Journal of Information Management 22(2): 127–142.

115.

Warnier

J-P

(1999) Construire la culture matérielle. L’homme qui pensait avec ses doigts. Paris: PUF.

116.

Wever

Del Castillo

(2006) Thinking out of the box: The unpacking experience of consumer electronics products. In: Proceedings of 15th IAPRI world conference on packaging, Tokyo, Japan, 4–5 October, pp. 136–140. Tokyo, Japan: IAPRI.

117.

White

Lin

Dahl

Ritchie

RJB

(2016) When do consumers avoid imperfections? Superficial packaging damage as a contamination cue. Journal of Marketing Research 53(1): 110–123.

118.

Wiedmann

K-P

Hennigs

Pankalla

Kassubek

Seegebarth

(2011) Adoption barriers and resistance to sustainable solutions in the automotive sector. Journal of Business Research 64(11): 1201–1206.

119.

Wolfinbarger

Gilly

(2001) Shopping online for freedom, control, and fun. California Management Review 43(2): 34–55.

120.

Yen

(2008) Factors influencing online auction repurchase intention. Internet Research 18(1): 7–25.

121.

Zeithaml

(1988) Consumer perceptions of price, quality, and value: A means-end model and synthesis of evidence. Journal of Marketing 52(3): 2–22.

122.

Zhao

Lynch

Chen

(2010) Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of Consumer Research 37(2): 197–206.