The moderating effect of temporal distance on perceived benefits and risks of online retail categories : a comparative study

(1)

The moderating effect of temporal distance on perceived benefits

and risks of online retail categories: a comparative study

Abstract

This study discusses the effect of online retail categories through perceived benefits and perceived risks on purchase intention, moderated by the variable of temporal distance. More specifically, the effects of online groceries versus online apparel on perceived benefits and perceived risks will be manipulated by temporal distance. Therefore this article measures main and interaction effects on purchase intention of 4 variables: (1) perceived benefits, (2) perceived risks, (3) online retail category, and (4) temporal distance. This study hypothesizes that online groceries have higher perceived risks and therefore lower purchase intention. Whereas online apparel has higher perceived benefits and therefore higher purchase intention. On the contrary, results indicate that online apparel has significantly less perceived benefits than online grocery. Whereas online groceries have significantly less perceived risks than online apparel. However, within the framework of this study, perceived benefits do have significant positive effect on purchase intention, whereas perceived risks have a significant negative effect on purchase intention. Temporal distance only had a significant positive moderating effect on perceived benefits.

Student: Martine Vogel Number: 10000884

Course: Thesis Final Draft Due Date: January 29th 2016 Supervisor: Alfred Zerres

Statement of Originality

This document is written by Student Martine Vogel, who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

(2)

1

List of Table and Figures

Tables

Table 1: Frequency Table ... 22

Table 2: Descriptive analysis with t-test on main and control variables ... 24

Table 3: Descriptive Table of Minimum, Maximum, Skewness and Kurtosis ... 25

Table 4: Bivariate analysis: means, standard deviations, correlations and reliabilities ... 26

Table 5: Overview of (un)supported hypotheses ... 27

Table 6: Moderated Mediation Regression (PROCESS) Results ... 28

Table 7: Conditional (Indirect) Effects of Online Retail Category (X) on Purchase Intention (Y) at values of Temporal Distance (W) ... 29

Table 8: T-test analysis on items of Perceived Benefits and risks at values of Online Retail Category ... 32

Figures Figure 1: Conceptual Model ... 17

(5)

4

1. Introduction

Most, if not all, consumers are familiar with the traditional, brick-and-mortar grocery shopping experience. Considering that groceries provide us with some of our basic physiological needs (Maslow, 1943), it is an essential and routine experience. According to Erasmus et al. (2014), groceries are the least complex of purchase decisions. Therefore, groceries as a general topic of interest, has triggered decades of research. Literature covers a plethora of topics related to grocery shopping, e.g. how shopping experience is effected by price knowledge (Jensen & Grunert, 2014), customer satisfaction (Esbjerg, et al., 2012), assortment (Kwak et al., 2015).

Despite the routine experience of grocery shopping as a whole, consumers do not take to the idea of shopping for groceries online. According to the AT Kearny Global Retail E-Commerce Index (Ben-Shabat et al., 2015) the lowest average global percentages for online purchases categories are groceries (45%) and household items (45%). The Centraal Bureau voor de Statistiek (CBS) (2015) states that Dutch consumers most frequently purchase (1) vacation paraphernalia and (2) clothing and sport articles online. One of the least purchased items out the 13 categories (ranked 5th from the bottom) are foodstuff. It would be of theoretical interest to understanding what was causing groceries to be unpopular as an online product category. The CBS (2015) states that 83% of Dutch people who are active on the interest, engage in online purchasing; an increase of 1% from 2013. If the number of active online shoppers is increasing in general, why are consumers, specifically, not shopping for groceries, despite the essentiality of foodstuff? Literature has attempted to answer the question of why grocery retail is lacking online popularity.

This study will begin by analyzing all relevant preceding literature to get an understanding of what has been previously researched on the topic of online retail. Based on the findings that online groceries lag behind other online retail categories, (CBS, 2015), it is of importance to understand why this occurs. Therefore, online groceries will be compared to online apparel retail, a relatively more successful online retail category. Moreover, purchase intention will be defined and its relation to purchase behavior will be explained. This study will then analyze how online groceries and online apparel differ in terms of purchase intention. The relationship between perceived benefits and perceived risks on purchase intention will then be examined. Again, online groceries and online apparel will be analyzed on their different and expected effects on the

(6)

5 perceived benefit and risk, and purchase intention relationship. Lastly, the concept of construal level theory, and more specifically, temporal distance, will be defined in light of this study.

According to Forsythe et al. (2006) and Chiang & Dholakia (2003) perceived benefits have a positive relationship with purchase intention. Moreover, if the online retail category is apparel, perceived benefits will increase in comparison to groceries, which in turn means that purchase intention will increase.

Furthermore, literature has validated a negative relationship between perceived risks and online shopping (usage) intention (Chu et al., 2010). This implies that when perceived risks are high, purchase intention is low. Considering that apparel is a successful retail category according to the CBS (2015), this study will assume that this is because apparel has a relatively lower level of perceived risks. Whereas when the online retail category is groceries, considering it is a relatively unsuccessful online retail category, (CBS, 2015), it will be assumed that groceries have a relatively higher level of perceived risks.

As was stated in the previous paragraph, online apparel is assumed to have higher perceived benefits. Which, based on Eyal et al.’s (2004) analysis, would assume that online apparel is an abstract phenomenon. This would imply that in the presence of high temporal distance (abstract) the relationship between online retail category, perceived benefits and risks, and purchase intention, will not be affected. However, in the presence of low temporal distance, it will be assumed that perceived risks increase and perceived benefits decrease.

Given that online groceries are concrete, low temporal distance should not have an effect the relationship between online retail category, perceived benefits and risks, and purchase intention. However, when high temporal distance (abstract) is introduced, the level of perceived risks should decrease, and perceived benefits increase. Which in turn ultimately increases purchase intention.

To analyse the data, standard SPSS instruments and techniques have been used. Variable descriptives are measured and analysed by generating means, standard deviations and the skewness and kurtosis. A Bivariate Correlation technique is used to measure relationships that exist between variables. A Moderated Mediation Regression is performed through the use of PROCESS by Hayes (2013).

(7)

6 Participants completed a 10-minute online survey, in the form of an experimental design. There were a total of 210 original respondents, which led ultimately to 150 valid cases.

This study finds, similar to previous studies, a significant, positive effect between perceived benefits and purchase intention. A significant, negative effect is found between perceived risks and purchase intention, similar to Jarvenpaa et al. (1999), Van der Heijden & Verhagen (2004), Park et al. (2005) and Chang & Chen (2008).

Moreover, this study finds that online retail category has a positive direct relationship with perceived risks instead of the expected negative direct relationship. Respondents indicated that not being able to touch or feel the product(s) before purchasing was more important in the apparel category than groceries. These findings could imply that apparel is indeed an experience good, as mentioned by Young Kim & Kim (2004).

This study also hypothesized that online retail category would have a positive effect on perceived benefits. In other words, the expectation was that online apparel retail would increase perceived benefits, and online grocery retail would decrease perceived benefits. Considering apparel’s success on the Dutch market and therefore higher purchase intention, it was assumed that the benefits for online apparel would be higher. However, this study shows contradicting results. The online grocery category had higher perceived benefits than online apparel. More specifically, respondents found that shopping in the privacy of their homes was more important when doing online groceries than online apparel. Moreover, respondents indicate that not having to wait to be served by a cashier or employee is more important when shopping for groceries than apparel.

There are several limitations to be mentioned in this study. The first limitation is in the sample size. Secondly, the generalizability of the research is limited to Dutch consumers. Thirdly, the manipulations for temporal distance may not have been extreme enough relative to the speed of consumerism. Lastly, the discussion section of this study highlights a potential disproportionate weight of products within online retail category.

2. Literature Review

This literature review will first analyse the current state of online retail. It will then analyze how purchase intention relates to online retail and the preceding literature on the topic. The

(8)

7 purchase intentions of online groceries and online apparel will then be covered briefly, determine the effect of online retail categories on purchase intention. Moreover, perceived benefits and Risks will be analyzed to understand how these factors play a role in the relationship between online retail categories and purchase intention. Finally, the concept of construal level theory will be introduced, more specifically temporal distance, to understand how to manipulate the above mentioned relationship.

2.1. Purchase intention

A consumer is a person who purchases goods and services for their own personal needs (Collins English Dictionary, 2016). Based on this definition of a consumer, it is clear that the goal of a consumer is to purchase a good or a service. Therefore, purchasing behavior can be seen as the ultimate indicator of consumer satisfaction. However, much of marketing research is performed under time and/or financial constraints, and therefore have a tendency to use purchase

intention instead of actual purchasing behavior (Hoch & Ha, 1986). The concept of purchase intention derives initially from the Theory of Reasoned Action (Fishbein & Ajzen, 1975) (Ajzen

& Fishbein, 1980), which states that the intention to perform a behavior is a good indicator of actual behavior. Sheppard et al. (1988) validated the significance of behavioral intention as an accurate indicator of behavior. The behavioral step in consumerism is essentially the purchasing of a good or service. Which in turn would imply that purchase behavior is preceded by purchase intention. Therefore this study will utilize purchase intention as a behavioral indicator, in line with similar articles (e.g. (Aghekyan-Simonian et al., 2012), (Park & Stoel, 2005), (Park et al., 2005).

As can be seen in AT Kearny Global Retail E-Commerce Index (Ben-Shabat et al., 2015) and the CBS (2015), inherent differences in purchase behavior exist between online retail categories. Chiang & Dholakia (2003) validated that, among other things, product types did, in fact, have a significant effect on online purchase intention. In other words, an apparent consensus among academics states that the online retail category will have a significant effect on a consumer’s online purchase intention. The large differences in the percentage of consumers active in a specific online retail product category, as stated by the CBS (2015), indicate that there is indeed a relationship between online retail category and purchase intention.

(9)

8

2.1.1. Online groceries

Verhoef & Langerak (2001) define online grocery shopping as the ability for consumers to electronically order groceries from their home environments, and in turn, receive these deliveries in their home environments. However, considering the current trend of mobile applications and mobile shopping, the “home” environment aspect of Verhoef & Langerak’s (2001) definition has also expanded to private, yet out-of-home environments. The channel touchpoints have expanded significantly since Verhoef & Langerak’s (2001) findings. Although relatively outdated, Raijas (1997) highlights that dissatisfaction among online grocery shoppers is higher relative to traditional grocery shoppers. If consumers are dissatisfied with a particular online retail category, they may refrain from purchasing from that category. Raijas (2002) further states that online grocery consumers find it difficult to search for products and are faced with uncertainty about the quality of those products. The author also states that consumer who engage in online grocery retail tend to buy products that are unperishable or heavy to carry. Which the author, in turn, counters with the finding that perishable or fresh products are tend to purchase in the brick-or-mortar grocery store.

Only a fraction of the literature is covered above, on the topic of online groceries, in order to give an idea of the what the “mood” of the findings on online grocery retail is. As can be retracted from the paragraph above, the category suffers from negative perceptions. Bhatnagar & Ghose (2004) highlight that the success of a particular online retail category depends on how the consumer perceives the category. As mentioned earlier, it will be assumed that the grocery category is relatively unsuccessful (CBS, 2015). Furthermore, based on Bhatnagar & Ghose’s (2004) study, it is assumed that this is due to the consumer’s perception of the category.

Therefore, we will assume that online groceries is unsuccessful, due to negative consumer perceptions which in turn result in low level of purchase intentions.

2.1.2. Online apparel

As mentioned, Chiang & Dholakia (2003) state that product types have a significant effect on online shopping intention. To ensure heightened heterogeneity in the results and a rich comparative analysis, this study will use apparel as the opposing online retail category. The CBS (2015) states that apparel is the second most frequently purchased online retail category in the Netherlands. This will allow for a comparison of purchase intention between online groceries and

(10)

9 online apparel. This comparative analysis derives primarily from Chiang & Dholakia’s (2003) study, which also compared two online product categories, i.e. books and perfume. The authors also classified the categories as two different types of “goods”, as defined by Nelson (1974). Books were classified as “search goods” and perfume was classified as an “experience good”. To better understand how these categorizations are relevant to the current study, it is important to discuss their role in the comparative analysis of online groceries and online apparel.

2.1.3. Search vs experience goods

Nelson (1974) classified goods, or products, into two groups known as search goods and experience goods. Search goods refer to products where all information is available prior to purchase. Whereas experience goods are products that require direct experience to determine the characteristics or information. In light of the current study, Nelson (1974) classifies men’s and women’s clothing as “search goods” and classifies groceries as non-durable “experience goods”. However, considering that online retail did not exist when Nelson (1974) published his work, the online equivalents of these product categories were not taken into consideration. If we take into consideration that groceries require physical inspection in both online and offline retail environments (Aghekyan-Simonian et al., 2012), this study will follow Nelson’s (1974) assumption that online groceries are experience goods.

Nelson (1974) classifies apparel as a search good in an offline context. However, apparel has been subject to opposing research. Bhatnagar & Ghose (2004) consider the reality of the Internet adding a different dimension to the traditional classification of search and experience goods. The authors claim that some categories have faced a “reclassification”. This could account for the discrepancy in the categorization of search goods and experience goods. Countering Nelson’s (1974) work, are Young Kim & Kim (2004), who claim that apparel is a category, which, in an online setting, is subject to a touch-and-feel environment; meaning it’s an experience good. However, considering heightened technological improvements, such as website design features (Hausman & Siekpe, 2009), consumers have access to all the information regarding apparel items prior to purchase. This means apparel may still be a search good. For the purpose of this study, it will be assumed that Nelson’s (1974) assumptions, that apparel is a search good, still holds today.

In order to understand the relevance of the search/experience good classification, it is important to highlight its relation to purchase intention. Chiang & Dholakia (2003) claim that the

(11)

10 intention to shop online is higher with search goods (apparel) than experience goods (groceries). This finding supports the conclusions of the CBS (2015) that state that apparel (search good) is more successful as an online product category than grocery (experience good) in terms of purchase behavior.

The classification of search goods and experience goods cannot fully explain the underlying drivers that differentiate the categories in purchase intention. However, it does give an indication of how groceries and apparel relate to purchase intention. Bhatnagar & Ghose (2004), however, as mentioned above, state that the success of an online retail category depends on the way in which the consumer perceives the category. A way in which consumers can perceive an online retail category is through weighing the pro’s, the reasons for doing something, and the con’s, the reasons for not doing something. A large fraction of online consumer behavior literature has been devoted to understanding these pro’s and con’s, also known as the benefits or risks of potential behavior, which drive consumers to purchase, or intend to purchase online (Verhoef & Langerak, 2001); (Bhatnagar & Ghose, 2004); (Bauer, 1960); (Aghekyan-Simonian et al., 2012).

2.2. Perceived benefits and risks

When the Internet was introduced as a purchasing medium, it triggered a peak in marketing and consumer literature. Because online consumer behavior became an exponentially popular subject in the 1990s and 2000s, the literature on the topic is extensive and overlapping. E.g. Verhoef & Langerak (2001) measured the relationship between advantages, or benefits, and disadvantages, or risks, of online grocery shopping e.g. physical effort, time pressure, shopping enjoyment, perceived complexity.

Considering the magnitude of academic literature on the topic of online retail, Chang et al. (2005) took the initiative to compile all the preceding literature that related to the topic of online consumer behavior. The authors compiled a total of 45 studies, characterized by 4 dependent variables, namely (1) intention and usage of online shopping, (2) attitude (3) risk perception, and (4) trust. Although this provided consumer behavior literature with an overview of all valuable insights, it did not succeed in developing a centralized framework to align the overriding factors that affect online purchase intention. Forsythe et al. (2006) tackled this lack of consensus by developing a framework that categorized the most salient variables under two primary factors, namely, perceived benefits and perceived risks.

(12)

11

2.2.1. Perceived benefits

Benefits can be defined as being the positive aspects of doing something, and in consumerism, the positive aspects of purchasing a good or service. Therefore, according to Bhatnagar & Ghose (2004) online shoppers aim to maximize their benefits and minimize their risks.

Davis (1989) initially developed a technologically-centered view on online consumer behavior. The author developed what is known as a Technology Acceptance Model (TAM), which states that consumers must first be mentally “willing” to use the technology, or online platform, before they engage in using the technology. Davis (1989) highlights that consumers develop mental responses, i.e. perceived usefulness and perceived ease of use, which ultimately results in a behavioral response, i.e. technology usage intention. Perceived usefulness (PU) is defined as the degree to which a person believes that using a particular system would enhance his or her job performance, or reaching one’s goal. Perceived ease-of-use (PEoU) refers to the degree to which a person believes that using a particular system would be free of effort. However, Davis’ (1989) technologically-centered view on consumer behavior, was also limited to analysis of usage intention, rather than the effects of purchase intention. Various authors filled this gap by analyzing the relationship between TAM and online purchase intention (Van der Heijden et al., 2003), (Moon & Kim, 2001).

Despite the extension of literature around TAM, Jarvenpaa & Todd (1997) criticized the lack of consumer-orientation, and the lack of “negative” factors, in the work of Davis (1989). Consequently, the authors were the first to develop a “customer-centered” view of online retail. The authors performed an experimental study to determine what the most salient factors in online shopping were from the consumers’ own perspective. Results showed that convenience was the most salient factor for consumers choosing to use an online retail medium. Bhatnagar & Ghose (2004) also identified convenience as a perceived benefit. However, they used convenience as the sole representative factor of perceived benefits. Furthermore, Jarvenpaa & Todd (1997) found that product perceptions, shopping experience and customer service are important positive factors in the consumer’s decision to use an online retail medium. Taking into consideration these factors, Forsythe et al. (2006), in conjunction with the work of Chiang & Dholakia (2003), validated the following online retail benefits: shopping convenience, product selection, ease/comfort of

(13)

12 shopping, and hedonic/enjoyment. These benefits are also part of this study. Although not identically worded, the perceived benefits of Jarvenpaa & Todd (1997) and Forsytheet al. 2006) can be seen as similar. Moreover, shopping convenience is in line with Davis’ (1989) PU, whereas the ease/comfort of shopping is in line with Davis’ (1989) PEoU.

Moreover, Chintagunta et al. (2012) argue that bulky and heavy retail categories tend to be the most frequently purchased online retail categories because of the shopping convenience benefits. Therefore, the more heavy or bulky groceries get, the more consumers are assumed to use an online grocery retail platform. According to Erasmus et al. (2014), (offline) groceries are the least complex of all (offline) product categories. The assumption in this study is that general online groceries are perceived as least complex, and not that heavy. Therefore, solely considering the effect of shopping convenience, it is expected that the perceived benefits will be low, which in turn would mean that the purchase intention decreases. On the contrary, apparel which will be assumed to be a bulky, will benefit from its convenience aspects, which in turn means that perceived benefits are relatively higher, and therefore purchase intention is higher. Based on the previous stated studies, e.g. Forsythe et al. (2006) and Chiang & Dholakia (2003), this study will assume that perceived benefits and purchase intention have a positive relationship. If perceived benefits increase, purchase intention increases.

2.2.2. Perceived risks

Bauer (1960) was the first to introduce the concept of perceived risk in consumer behavior literature. Taylor (1974) defines perceived risk as the level of uncertainty a consumer feels it needs to cope with in a purchase situation. Perceived risk according to Coleman (1994) is a consumers assessment of the potential positive and negative outcomes of a purchase situation. Regarding online retail, consumers have higher levels of perceived risk when shopping on the Internet (Bhatnagar & Ghose, 2004). perceived risks was an aspect that Davis (1989) failed to

incorporate in the TAM. Featherman & Pavlou (2003) follow-up on the original TAM by empirically validating the influence of perceived risk on usage intention in electronic commerce. The authors state that perceived risk is a prominent barrier to consumer acceptance of an online retail platform. From the consumer-centric definition, Jarvenpaa & Todd (1997) state that online consumer intention is affected by the amount of risk that is attached to the experience. The authors found that the four following dimensions of consumer risk were most salient: economic, social,

(14)

13 performance, personal and privacy risk. Economic risk is related to the actual monetary loss associated with the purchase of a product. Social risk will be disregarded in this study, as is similar to (Forsythe et al.’s (2006) study. Performance risk is the risk of the purchased product not complying to the consumers’ expectations. This definition of performance risk is in line with the definition of Aghekyan-Simonian et al. (2012), who state that the inability to physically examine products, in online shopping environments, increases the consumer’s risk perceptions as they cannot touch, feel, or try on products prior to purchase. Personal risk is the consumer’s fear that something bad will happen to the consumer when using the product or during the shopping journey e.g. credit card information loss.

Literature has validated a negative relationship between perceived risks and online shopping (usage) intention (Jarvenpaa et al., 1999), (Van der Heijden & Verhagen, 2004) (Park et al., 2005) (Chang & Chen, 2008). This study will therefore assume that perceived risks and purchase intention have a negative relationship. If perceived risks increase the purchase intention decreases, and vice versa.

2.3. Construal level theory

Ho et al. (2015) developed a conceptual model that measured the (extended) TAM which includes perceived risk. This model is also used in the Featherman & Pavlou’s (2003) study. Ho et al. (2015) measure the effect of TAM and perceived risk on the usage intention of e-learning systems. The study analyzes online services, and not online retail platforms selling tangible products. However, the conceptual model is comparable to Forsythe et al.’s (2006) perceived benefits and risks framework. One specific difference is Ho et al.’s (2015) usage of a moderating factor called construal level theory (CLT). In the scope of this literature review, the variable of construal level as a moderating effect has not previously been applied to Forsythe et al.’s (2006) framework. The following paragraphs will define construal level theory, its relation to psychological and temporal distance, and its relevance to this study.

Trope et al. (2007) define CLT as a framework that aims to understand the level of abstraction and distance involved in consumers’ evaluation of objects or events. CLT defines these evaluations as mental representations, or mental construals, of current, past and future actions. In terms of online consumer behavior, CLT suggests that consumers develop mental construals of their decisions to purchase online. Low-level construals, refer to events that happen in close

(15)

14 proximity to the consumer, which are in turn perceived as concrete. On the contrary, high-level construals, refer to events that happen in further proximity to the consumer, which are perceived as abstract. The concept of a concrete event derives from the idea that events that take place in close proximity are vividly detailed with secondary features. On the other hand, an abstract event derives from the idea that events in further proximity rely primarily on the primary features.

2.3.1. Psychological Distance

Trope et al. (2007) define CLT as being a framework that combines the concepts of abstraction and distance. The concept of abstraction has been briefly defined in the previous paragraph. This paragraph will come to understand the concept of distance. Trope et al. (2007) define the concept of distance specifically as psychological distance. The authors distinguish 4 types of psychological distance, namely, social distance, hypothetical distance, spatial distance and temporal distance. Social distance refers to how much other people are involved or uninvolved in an event, or, for consumers, a purchase situation. For example, if one decides to purchase healthcare, a father might consider his family’s needs (low social distance, concrete), whereas a single man only needs to consider himself (high social distance, abstract). Hypothetical distance refers to the likelihood of something occurring, including the uncertainty of consequences and outcomes. Hypothetical distance is high in a purchase situation such as a investing in a wedding, because the outcome of the wedding day itself is not clear. Low hypothetical distance could be experienced when a consumer purchases the same pack of cigarettes; the expectations tend to already be met. Spatial distance, is the physical distance between the e.g. the consumer and a future event. The distance between going to outlet stores, 3 hours away, and a regular store around the corner. Lastly, and most importantly for this study, is the concept of temporal distance. temporal distance is defined as the amount of time that exists relative to the event. For example, if a consumer wants to purchase something next year, temporal distance is high, which also implies that abstraction high, because a year is a long time relative to in comparison to next week.

2.3.2. Temporal distance

Eyal et al. (2004) specifically validated the effect of temporal (time) distance on the perceived weight of pros versus cons. Pros and cons are the favorable and unfavorable reasons, respectively, for doing something (Collins English Dictionary, 2016). Pros and cons can also be referred to as advantages and disadvantages, as is similar to Verhoef & Langerak (2001).

(16)

15 Therefore, this study will assume that pros and cons are the perceived benefits and perceived risks, respectively.

Furthermore, according to Eyal et al. (2004) pros are more dominant in an abstract state-of-mind, whereas cons are more dominant in a concrete state-of-mind. This means that in an online retail environment, a concrete temporal mindset will result in a higher number of risks versus benefits, whereas an abstract temporal mindset will result in a higher number of benefits versus risks. In turn, as mentioned previously in this literature review, a larger amount of risks will have an negative effect on purchase intention, whereas a more benefits will have a positive effect on purchase intention.

Considering the non-complex (Erasmus et al., 2014) and unsuccessful (CBS, 2015) nature of (online) groceries, it will be assumed that online groceries will be perceived as a concrete phenomenon. This assumption is made because online groceries are expected to have low purchase intention, which is assumed to be because of the negative relationship between perceived risks and purchase intention. Given that Eyal et al. (2004) indicate that low temporal distance (concrete) increases the level of perceived risks, online groceries will be assumed to be a concrete phenomenon. However, when temporal distance is added into the equation, it may be possible to manipulate the relationship between perceived benefits and risks on purchase intention. Given the assumption that online groceries are concrete, low temporal distance (concrete) is expected not to manipulate the relationship between online retail category, perceived benefits and risks, and purchase intention. However, when high temporal distance (abstract) is introduced, the level of perceived risks should decrease, and perceived benefits increase. Which in turn ultimately increases purchase intention.

Online apparel, considering its successful nature, is assumed to have higher perceived benefits. Which, based on Eyal et al.’s (2004) analysis on pros and cons, would assume that online apparel is an abstract phenomenon. This would imply that in the presence of high temporal distance (abstract) the relationship between online retail category, perceived benefits and risks, and purchase intention, will not be manipulated. However, in the presence of low temporal distance, it will be assumed that perceived risks increase and perceived benefits decrease.

(17)

16 2.4. Hypotheses

Within this study, independent variable online retail category is categorized as a dummy variable, where 0 represents online groceries and 1 represents online apparel. The moderating variable temporal distance is also indicated as a dummy variable, where low temporal distance will be categorized as 0 and high temporal distance will be categorized as 1.

As previously stated, according to Forsythe et al. (2006) and Chiang & Dholakia (2003) perceived benefits have a positive relationship with purchase intention. Moreover, online retail category has a positive relationship with perceived benefits. If the online retail category is apparel, perceived benefits will increase, which in turn means that purchase intention will increase. If the online retail category is groceries, then online perceived benefits will decrease, and purchase intention will decrease. Therefore:

H1: Online retail category has a positive, indirect relationship with purchase intention, mediated

by perceived benefits.

As previously mentioned, literature has validated a negative relationship between perceived risks and online shopping (usage) intention (Chu et al., 2010). This implies that when perceived risks are high, purchase intention is low. Considering that apparel is a successful retail category according to the CBS (2015), this study will assume that this is because apparel has a relatively lower level of perceived risks. Whereas when the online retail category is groceries, considering it is a relatively unsuccessful online retail category, (CBS, 2015), it will be assumed that groceries have a relatively higher level of perceived risks. Therefore:

H2: Online retail category has a negative, indirect relationship with purchase intention mediated

by perceived risks.

H3: Online retail category has a negative direct relationship with perceived risks. H4: Online retail category has a positive relationship with perceived benefits.

As was stated in the previous section, online apparel is assumed to have higher perceived benefits. Which, based on Eyal et al.’s (2004) analysis, would assume that online apparel is an abstract phenomenon. This would imply that in the presence of high temporal distance (abstract) the relationship between online retail category, perceived benefits and risks, and purchase

(18)

17 intention, will not be affected. However, in the presence of low temporal distance, it will be assumed that perceived risks increase and perceived benefits decrease.

Given that online groceries are concrete, low temporal distance should not have an effect the relationship between online retail category, perceived benefits and risks, and purchase intention. However, when high temporal distance (abstract) is introduced, the level of perceived risks should decrease, and perceived benefits increase. Which in turn ultimately increases purchase intention.

H5: Temporal distance has a negative moderating effect on the relationship between online retail

category on perceived risks.

H6: Temporal distance has a positive moderating effect on the relationship between online retail

category on perceived benefits.

2.4.1. Research Model

The visualization of the research model used in this study can be found in Figure 1. The expected relationships and hypotheses can be found along the relationship arrows. Each variable is presented with an additional letter. These letters derive from the research models found in Hayes’ (2013) PROCESS. X refers to the dependent variable. Y refers to the dependent variable. M1 and M2 refer to the mediating variables. W refers to the moderating variable.

Figure 1: Conceptual Model

Online Retail Category (X)

Perceived Benefits (M1)

Purchase Intention (Y)

Perceived Risks (M2) Temporal Distance (W) + -+ + -H1 H2 H3 H4 H5 H6 Control variables: Age Education Gender Frequency of Purchase Previous Experience

(19)

18

3. Methodology and data

3.1. Procedure and materials

3.1.1. Pilot survey

A pilot survey was distributed to a total of 25 respondents. The pilot survey had the goal of filtering out flaws in the structure, language use and question formation. The respondents were asked to give feedback on any glitches in the survey. These comments were taken into consideration when developing the final draft of the survey. Some feedback included topics such as faulty question redirection, spelling errors, and vagueness in questions. The respondents who participated in the pilot survey were not asked to participate in the final survey; this to avoid pre-exposure bias or carry over effects.

3.1.2. Final Survey

Participants completed a 10-minute online survey, in the form of an experimental design (see Appendix 7.1). Respondents were first introduced to the study, and its relation to the Amsterdam Business School. The study had the option of being completed in English and Dutch. Considering that majority of the respondents were native-Dutch speakers, it was important that language was not a barrier during this survey. The Dutch content was further translated and edited by a nativDutch speaker prior to distribution. Respondents were gathered through social media platforms, e-mail and Whatsapp. Convenience sampling and snowball sampling were used as sampling techniques. Therefore the initial sampling was a non-probability sample. However, due to the experimental design, and randomization process, the final sample was exposed to a probability sampling technique. Respondents were randomly assigned to one of the four conditions. This randomization was performed by Qualitrics itself to ensure that there was no selection bias. The manipulation text is as follows:

Conditions: “It's been a while since you've thrown a good party so you're considering throwing a party (next weekend/March 2016). You want to make sure that your friends and family have the best time they can at your party. Currently, you are considering buying (groceries/an outfit)for the party.

As can be seen above, in this study it is assumed that low Temporal Distance can measured by “next weekend” and high Temporal Distance can be measured by “March 2016”. The survey was distributed in December 2015. March 2016 was chosen as the condition with high Temporal

(20)

19 Distance as it was, more or less, three months in advance. March is also a month where no nation-wide holidays are celebrated and therefore the respondents were given no biased occasion to refer to.

The Online Retail Category, was either online groceries or online apparel. Respondents were randomly assigned to each of the categories.

The mediating variables and dependent variable were measured on 7-point Likert scales. All specific questions used in the online survey can be found in Appendix 7.1. Respondents were first asked to answer 3 questions on the dependent variable, Purchase Intention. These variables derive from the methodological framework validated by Hausman & Siekpe (2009) and Poddar et al., (2009). One question was reversed to ensure that respondents remained attentive.

Respondents were then asked to answer 8 questions relating to Perceived Benefits, in the form of a question matrix. All questions were chosen from the validated Perceived Benefits constructs in Forsythe et al.’s (2006) study. Not all of Forsythe et al.’s (2006) questions were used as some questions were obsolete or irrelevant for this study. As mentioned in previous sections, Forsythe et al. (2006) also classified all the questions under constructs. Although these sub-constructs were not taken into consideration during this study, they are theoretically still present, and will therefore be discussed. The Perceived Benefits in this study are categorized as follows: shopping convenience (PB_1, PB_2, PB_3), product selection (PB_4, PB_5), ease of shopping (PB_6, PB_7), enjoyment/hedonic (PB_8).

Respondents were then asked 8 questions relating to Perceived Risks. These questions also derive from Forsythe et al. (2006), where some questions were, again, omitted due to obsolescence or irrelevance. The questions related to the following sub-constructs: financial risk (PR_1, PR_2, PR_3, PR_4), product risk (PR_5, PR_6), and time risk (PR_7, PR_8).

The respondents were frequently reminded of what their manipulations as they were presented above every new set of questions. This helped to ensure that the manipulation resonated in the respondents mind when answering the questions. To ensure that the manipulations were successful, respondents were asked to answer the following two questions: (1) What was it, according to this survey, that you were buying for your party, and (2) when was, according to this

(21)

20 survey, your party going to take place. These manipulation checks were free-fill questions and ensured that respondents had to reflect on the answer from their own memory.

The control variables were based on the control variables in Ho et al.’s (2015) study. The first question was whether respondents live in the Netherlands. This question controlled for any respondents that were not Dutch consumers, or consumers living the Netherlands. The other control variables are Age, Gender, Education, Previous Experience, Frequency of Purchase and Previous Experience within Online Retail Category. All control variables will be listed with their respective SPSS code in brackets and italics. Age (age) is a ordinal variable, consisting of 8 categories based on a frequently used categorization designed by Qualtrics. Gender (d_gender) is considered a nominal (dummy) variable consisting of 2 categories, male (0) or female (1). Education (edu) is an ordinal variable, consisting of 8 categories. Previous Experience (d_prepur) is a nominal variable (No = 0, Yes = 1) asking whether respondents have previously purchased anything (good or service) online. Frequency of Purchase (freqpur) is a measured on a 5-point Likert scale. Respondents were asked to identify the frequency of their online purchase behavior. Lastly, respondents were asked whether they had purchased anything within their Online Retail Category. For example, respondents in the grocery condition were asked whether they ever purchase groceries online. Whereas, respondents in the apparel condition were asked whether they ever bought apparel (an outfit) online.

3.2. Frequencies

There were a total of 210 original cases. All cases with missing values, defined by whether the respondent completed the survey, were filtered out of the dataset (38 cases, leaving 172 valid cases). Furthermore, all manipulation check answers were manually compared to the condition the respondents were given. If the answers given were not aligned with the respective condition, the case was filtered out, as the respondent’s manipulation did not work, and therefore cannot provide valid answers. These failed-manipulation cases were a total of 16 cases which leaves a total of 156 respondents or valid cases. Lastly, 6 cases had extreme answers when the reverse question (rPI3) was reversed accordingly (PI3). This means that these questions were not answered correctly and would ensure false representation. They were therefore removed. The sample is therefore based

(22)

21 on 150 valid cases. All the frequencies can be found in Table 1. The frequencies will be discussed in full in the order in which the Table 1 is arranged.

Respondents were also asked whether they had previously purchased anything in their given condition. For example, respondents who were categorized under “apparel” were asked whether they had previously bought clothing online; respondents categorized under “groceries” were asked whether they had previously bought groceries online. The responses were as follows: No (38.7%) and Yes (61.3%). It is important to understand how these answers are distributed among the categories. A total of 70 respondents were assigned to the grocery condition and 80 respondents to the apparel condition. Within the grocery condition 46 (65.7%) respondents replied “No” and 24 (34.3%) respondents replied “Yes”. Within the apparel condition, 11 (14%) replied “No” and 69 (86) replied “Yes”. As can be assumed from this information, the experience of online retail, with this study’s respondents, is primarily within apparel, and less so with grocery. Gender was unequally distribution. The gender split was male (38%) female is (62%). This could be due to the removal of missing values and the manipulation check. There were a total of 8 Age categories (Less than 16, 16 to 19, 20 to 24, 25 to 34, 35 to 44, 45 to 54, 55 to 64, 65 or over) from which respondents could choose from. There were no respondents under the age of 16. Furthermore, when age is put into a histogram, it can be seen that the distribution is relatively normal (see Appendix 7.2397.1). The majority of respondents fall between 20 and 44 (three middle-range age categories). The ages are distributed as follows: 16 to 19 (6%), 20 to 24 (19.3%), 25 to 34 (38%), 35 to 44 (17.3%), 45 to 54 (9.3%), 55 to 64 (8.7%), 65 or over (1.3%).

Education consisted of 8 categories (Less than High School, High School, MBO Degree, HBO Bachelors Degree, HBO Masters Degree, University Bachelors Degree, University Masters Degree, Doctoral Degree (PhD)), from which respondents could choose a category based on their highest obtained degree. There were no respondents in the categories “Less than Highschool” and “Doctoral Degree (PhD)”. Education, when put into a histogram, is not normally distributed. Most respondents (40.7%) are University Masters Degree Graduates. Followed by HBO Bachelor’s Degree (22%), High School (12.7%), MBO Degree (10%), University Bachelors Degree (7.3%), HBO Masters Degree (7.3%).

Respondents were asked whether they had previously purchased a good or a service online. Only 1 respondent replied “No” and the other 149 respondents replied “Yes”. Respondents were

(23)

22 then asked how frequently they purchase something (good or service) online. Respondents were asked to choose from 5 categories (Never, Rarely, Sometimes, Often, All the Time). The one respondent that had never bought anything online, was, as expected, the only respondent to choose “Never”. The distribution of frequency of purchase was quite normal. The distribution across categories is as follows: Never (0.7%), Rarely (7.3%), Sometimes (34%), Often (56.7%), and All of the Time (1.3%).

As was stated in the Methodology and data section of this study, respondents were randomly assigned to one of two types of Temporal Distance, namely, low Temporal Distance and high Temporal Distance. Respondents were categorized relatively equally, namely, low Temporal Distance had 73 respondents (48.7%) and high Temporal Distance had 77 (51.3%) respondents.

Respondents were also randomly assigned to an Online Retail Category, namely online groceries (48.7%) or online apparel (53.3%). As can be seen, the randomization also ensured for a relatively equal division of respondents amongst the two groups.

Table 1: Frequency Table

Item Categories % of respondents (n) Item Categories % of respondents (n)

Gender Male 38% (57) Never 0.7% (1)

Female 62% (93) Rarely 7.3% (11)

Sometimes 34% (51)

Age Less than 16 0% (0) Often 56.7% (85)

16 to 19 6% (9) All of the Time 1.3 % (2)

20 to 24 19.3% (29) 25 to 34 38% (57) No 38.7% (58) 35 to 44 17.3% (26) Yes 61.3% (92) 45 to 54 9.3% (14) 55 to 64 8.7% (13) Groceries 65 or over 1.3% (2) No 66% (46) Yes 34% (24)

Education Less than Highschool 0% (0)

Highschool 12.7% (19) Apparel

MBO Degree 10% (15) No 14% (11)

HBO Bachelors Degree 22% (33) Yes 86% (69)

HBO Masters Degree 7.3% (11)

University Bachelors Degree 7.3% (11) Product Category Groceries 46.7% (70)

University Masters Degree 40.7% (61) Apparel 53.3% (80)

PhD 0% (0)

Temporal Distance Low 48.7% (73)

No 0.7% (1) High 51.3% (77) Yes 99.3% (149) Frequency of Online Purchase Previous Purchase Previous Purchase Within Retail Category Total

Previous Purchase Within Retail Category

(24)

23 3.3. Reliability Analysis

The current study has a total of 19 measured items. These items include Purchase Intention with 3 items (PI_1, PI_2, PI_3), Perceived Benefits with 8 items (PB_1, PB_2, PB_3, PB_4, PB_5, PB_6, PB_7, PB_8), and Perceived Risks (PR_1, PR_2, PR_3, PR_4, PR_5, PR_6, PR_7, PR_8). The complete measurement scale has a Cronbach’s Alpha of 0.746, which means that reliability is high. However, when each variable is measured separately, the reliability results vary. Purchase Intention has high reliability (α = 0.911). All Corrected-Item Total Correlation (CITCs) are very high (range = 0.814 to 0.854) and Cronbach’s Alpha if Item Deleted (CAID) are < ∆ 0.10. The reliability of Perceived Benefits is moderately high (α = 0.734). CITCs are not great, but acceptable. One item, however has a CITC of 0.299, which is 0.001 under the cut-off. However, the all CAIDs are < ∆ 0.10, and therefore there is no support for the deletion of this item. The reliability of Perceived Risks is moderately high (α = 0.762). Again, CITCs are not great, but they are acceptable. Again, one item, has a CITC of 0.294, which is 0.006 under the cut-off. However, all of the CITDs are < ∆ 0.10, and therefore there is no support for the deletion of this item.

4. Results

To analyse the data mentioned in the Methodology and data section, standard SPSS instruments and techniques have been used. Firstly, variable descriptives are measured and analysed by generating means, standard deviations and the skewness and kurtosis. The skewness and kurtosis of the variables give an indication of whether data is normally distributed. Secondly, the a Bivariate Correlation technique is used to measure relationships exist between variables. However, a bivariate correlation cannot provide insights into causal relationships between variables. Therefore, a Moderated Mediation Regression is performed through the use of PROCESS by Hayes (2013).

4.1. Descriptives

Table 2 provides an overview of all the means divided amongst the two online retail categories, and amongst the two Temporal Distance categories. A series of independent t-tests were done to analyse the variance between the means. The difference between the means are stated under “Delta”. Only a few of the differences in means were significant. Within the variable of Purchase Intention, the difference between online retail categories is significant (t = -4.107, p <

(25)

24 0.01). Apparel is therefore significantly different, and in this case higher, than the grocery mean. This is in line with CBS’ (2015) conclusions that online apparel retail ranked higher in popularity and purchasing behaviour amongst Dutch consumers than online groceries. These findings are comparable to H1 and H2’s expectations that Online Retail Category, although indirect, have a positive relationship with Purchase Intention.

Moreover, the mean difference between online retail categories and Previous Purchase within Retail Category is significant (t = -7.949, p < 0.01). These findings are in line with the findings in the Frequency section of this study. More specifically, the frequencies of Previous Purchase within Retail Category indicated that there was higher level of respondents that had previously purchased apparel than those that had grocery.

Within Temporal Distance, only one variable had a significant difference in means, namely Perceived Benefits (t = 3.327, p < 0.01). Low Temporal Distance had a higher average Perceived Benefits score (4.51) than high Temporal Distance (3.97). The expectation of H5 and H6, according to Eyal et al. (2004), was that low Temporal Distance would yield lower Perceived Benefits and high Temporal Distance would yield higher Perceived Benefits. However, the means are significantly opposed in comparison to the expectations. Low Temporal Distance has resulted in higher Perceived Benefits and high Temporal Distance has resulted in low Perceived Benefits.

Table 2: Descriptive analysis with t-test on main and control variables

The skewness for all the variables were within the -1 and +1 range (-0.921 to -.566). However, the kurtosis for PI_TOT (-1.395) is notable. Because the skewness of PI_TOT does not

M SD Grocery Apparel Delta t-test* Low High Delta t-test

Main variables Purchase Intention 3.15 1.69 2.58 3.65 -1.08 -4.107** 2.98 3.32 -0.34 -1.232 Perceived Benefits 4.24 1.03 4.33 4.15 0.18 1.039 4.51 3.97 0.54 3.327** Perceived Risks 4.73 1.01 4.60 4.84 -0.24 -1.474 4.74 4.72 0.01 0.084 Control variables Age 4.36 1.36 4.40 4.33 0.08 0.335 4.29 4.43 -0.14 -0.632 Gender 0.62 0.49 0.66 0.59 0.07 0.873 0.62 0.62 -0.01 -0.087 Education 5.09 1.86 5.26 4.94 0.32 1.048 5.07 5.10 -0.04 -0.116

Frequency of Online Purchase 3.51 0.68 3.50 3.51 -0.01 -0.111 3.49 3.52 -0.03 -0.235

Previous Purchase Retail Category 0.61 0.49 0.33 0.86 -0.53 -7.949** 0.63 0.60 0.03 0.409

* t-value in t-test for Equality of M eans ** significant at the 0.01 level (2-tailed)

Online Retail Category Temporal Distance

(26)

25 show a relatively high value, variable transformation is not needed. The detailed results for skewness and kurtosis can be found in Table 3.

Table 3: Descriptive Table of Minimum, Maximum, Skewness and Kurtosis

4.2. Bivariate Correlation

The bivariate correlations of the main variables in this study will be discussed first. The analysis for the control variables will follow. All the correlations can be found in Table 4. Only the significant correlations will be mentioned.

Purchase Intention has three highly significant, positive correlations with frequency of online purchase (0.263), previous purchase within product category (0.422) and Online Retail Category (0.320). This correlation provides insights relating to the direct effect between Online Retail Category (X) and Purchase Intention (Y), namely a positive relationship between Online Retail Category (apparel) and Purchase Intention.

Perceived Benefits has a significant, negative correlation with Age (-0.183) and Education (-0.168). It also has a significant, positive correlation with Frequency of Online Purchase (0.201) and Purchase Intention (0.184), and a highly significant, negative correlation with Temporal Distance (-0.264). This correlation implies a relationship between Perceived Benefits (mediator 1: M1) and Purchase Intention (Y), namely, when Perceived Benefits increase, Purchase Intention increases. This relationship is in line with H1. The negative correlation between Perceived Benefits and Temporal Distance implies that when Perceived Benefits increase, Temporal Distance decreases.

N M inimum M aximum M ean

Std.

Deviation Statistic Std. Error Statistic Std. Error

Main variables

Purchase Intention 150 1.0 6.7 3.151 1.6867 .429 .198 -1.181 .394

Perceived Benefits 150 1.9 6.9 4.235 1.0322 -.112 .198 -.272 .394

Perceived Risks 150 1.5 6.6 4.729 1.0079 -.548 .198 .149 .394

Online Retail Category 150 0 1 .53 .501 -.135 .198 -2.009 .394

Temporal Distance 150 0 1 .51 .501 -.054 .198 -2.024 .394

Control variables

Age 150 2 8 4.36 1.362 .566 .198 -.085 .394

Gender 150 0 1 .62 .487 -.499 .198 -1.774 .394

Education 150 2 7 5.09 1.864 -.317 .198 -1.395 .394

Frequency of Online Purchase 150 1 5 3.51 .683 -.921 .198 .617 .394

Previous Purchase Retail Category 150 0 1 .61 .489 -.470 .198 -1.803 .394

(27)

26 Perceived Risks have a highly significant, negative correlation with Education (-0.271). It has a significant, negative correlation with Purchase Intention (-0.193) and a highly significant, positive correlation with Perceived Benefits (0.249). The prior correlation implies that when Perceived Risks increase, Purchase Intention decreases. This finding is in line with the expectations of H2. An interesting finding is that when Perceived Risks increase, Perceived Benefits also increase. Online Retail Category has a highly significant, positive, moderate correlation (0.547) with Previous Purchase within Retail Category. There is a highly significant, negative weak correlation (-0.244) between frequency of online purchasing and gender. Previous Purchase within Retail Category has a highly significant, positive correlation (0.249).

Table 4: Bivariate analysis: means, standard deviations, correlations and reliabilities

4.3. Moderated Mediation Regression

To measure the causal relationship between the 5 variables presented in this study, a moderated mediation regression (Model 7) was performed through PROCESS (Hayes, 2013). Table 1 provides all the supported/unsupported hypotheses.

Results indicate that the conceptual model is highly significant (R² = 0.3617, p = 0.0000). Perceived Benefits had a highly significant, positive effect on Purchase Intentions (β = 0.3719, p = 0.0028). Therefore, H1 is supported. Perceived Risks has a highly significant, negative effect on Purchase Intention (β = -0.4776, p = 0.0001). Therefore, H2 is supported. Product category has a

M SD 1 2 3 4 5 6 7 8 9 10

Control variables

1. Age 4.36 1.36 --2. Gender 0.62 0.49 -.106 --3. Education 5.09 1.86 .014 -.104 --4. Frequency of Online Purchase 3.51 0.68 .026 -.224** .150 --5. Previous Purchase Retail Category 0.61 0.49 -.052 -.058 .067 .249**

--Main variables

6. Online Retail Category 0.53 0.50 -.028 -.072 -.086 .009 .547** --7. Temporal Distance 0.51 0.50 .052 .007 .010 .019 -.034 -.055

--8. Purchase Intention 3.15 1.69 -.083 .133 .011 .263** .422** .320** .101 (0.91)† 9. Perceived Benefits _{4.24 1.03 -.183}* _.105

-.168* .201* -.086 -.085 -.264** .184* (0.73)† 10. Perceived Risks 4.73 1.01 -.012 .074 -.271** -.003 -.073 .120 -.007 -.193* .249** (0.76)† * - Correlation is significant at the 0.05 level (2-tailed).

** - Correlation is significant at the 0.01 level (2-tailed). † - Cronbach's alpha

(28)

27 significant, positive effect on Perceived Risks (β.= 0.5881, p = 0.0304). This relationship implies that in the presence of groceries (dummy variable = 0) decrease Perceived Risks, whereas apparel (dummy variable = 1) increases Perceived Risks. The results are contrary to the assumptions made, which in turn means that H3 is not supported.

Results indicate that product category has a significant, negative effect on Perceived Benefits (β = -0.5057, p = 0.0476). This implies that grocery category (dummy variable = 0) increases perceives benefits, whereas apparel (dummy variable = 1) decreases the Perceived Benefits. Therefore, H4 is not supported. There was no significant interaction (conditional) effect between Temporal Distance, product category, and Perceived Risks. Therefore, H5 is not supported.

A significant, positive interaction (conditional) effect exists between Temporal Distance, product category, and Perceived Benefits (β = 0.7677, p = 0.0155). This implies that the moderation effect, as a whole, is present. However, when the relationship is closely examined, only the interaction between Perceived Benefits and low Temporal Distance has a significant effect (β = 0.0974, BootLLCI -0.4696, BootULCI -0.0213). Therefore, H6 is partially supported. These causal findings are in line with the test performed in the descriptives section of this study. The t-test results state that only the difference between means of Temporal Distance of Perceived Benefits (t = 3.327, p < 0.01) was significant.

Table 5 provides an overview of each of the hypotheses and whether these are supported or unsupported.

Table 5: Overview of (un)supported hypotheses

Figure 2: Moderated Mediation Regression (PROCESS) Conceptual Model with Results provides the same conceptual model as Figure 1: Conceptual Model but with the specific (significant) effects. However, it does not indicate whether these effects were in line with

Hypothesis (Not) Supported Description

H1 Supported Online retail category has a positive, indirect relationship with purchase intention, mediated by perceived benefits H2 Supported Online retail category has a negative, indirect relationship with purchase intention mediated by perceived risks H3 Not supported Online retail category has a negative direct relationship with perceived risks

H4 Not supported Online retail category has a positive relationship with perceived benefits

H5 Not supported Temporal distance has a negative moderating effect on the relationship between online retail category on perceived risks H6 Partially supported Temporal distance has a positive moderating effect on the relationship between online retail category on perceived benefits

(29)

28 hypotheses. Therefore, Table 6 provides a full review of all the effects (coefficients), significance levels (p) and standard errors (SE), and the original theoretical prediction signs (T.P.S.).

Figure 2: Moderated Mediation Regression (PROCESS) Conceptual Model with Results

Table 6: Moderated Mediation Regression (PROCESS) Results

Table 7 provides a thorough depiction of the specific conditional (indirect) effects that occur between the variables Online Retail Category, Perceived Benefits and Risks, and Temporal Distance. As was stated above, the interaction effect involving Perceived Risks was not significant and is therefore not analysed.

Online Retail Category (X)

Perceived Benefits (M1)

Purchase Intention (Y)

Control variables:

Age Education

Gender Previous Experience Previous Experience Within Category

Perceived Risks (M2) Temporal Distance (W) H1: 0.37 H2: -0.48 H3: 0.59 H4: -0.51 H5: Not significant H6: 0.77

Antecedent T.P.S.* Coeff. SE p T.P.S. Coeff. SE p T.P.S. Coeff. SE p

Main variables

Online Retail Category (X) + -0.506 0.253 0.048 - 0.588 0.269 0.030 + 0.835 0.283 0.004

Perceived Benefits (M1) --- --- --- --- --- --- + 0.372 0.122 0.003 Perceived Risks (M2) --- --- --- --- --- --- - -0.478 0.122 0.000 Temporal Distance (W) + -0.954 0.223 0.000 - 0.161 0.238 0.500 --- --- ---ORC x TD (XW) + 0.768 0.313 0.016 + -0.310 0.333 0.354 --- --- ---Control variables Gender 0.219 0.161 0.177 0.172 0.172 0.318 0.705 0.244 0.005

Prev. Purchase Retail Category -0.278 0.194 0.204 -0.409 0.206 0.049 0.855 0.297 0.005

Education -0.107 0.041 0.010 -0.133 0.044 0.003 -0.030 0.065 0.065

Age -0.153 0.057 0.008 0.004 0.060 0.943 -0.011 0.086 0.899

Frequency of Online Purchase 0.394 0.121 0.002 0.168 0.129 0.196 0.503 0.187 0.008

constant 4.638 0.596 0.000 4.629 0.633 0.000 0.865 1.065 0.418

* Theoretical prediction sign

R² = 0.2544 R² = 0.1166 R² = 0.3617

F = 6.0123, p = 0.0000 F = 2.3254, p = 0.0225 F = 9.9856, p = 0.0000 Perceived Benefits (M1) Perceived Risks (M2) Purchase Intention (Y)

(30)

29 Table 7: Conditional (Indirect) Effects of Online Retail Category (X) on Purchase Intention (Y) at values of Temporal Distance (W)

5. Conclusion

5.1. Discussion

Although the direct effect of Online Product Category and Purchase Intentions was not explicitaly measured in this study, there was a moderately significant positive effect present (β = 0.8646, p = 0.0037). This causal relationship implies that the Purchase Intention is higher for online apparel retail than for online groceries. This finding supports the initial assumption that online grocery retail had a lower Purchase Intention than online apparel retail (CBS, 2015).

Moreover, similar to previous studies, (Forsythe et al., 2006) and (Chiang & Dholakia, 2003), a significant, positive effect was found between Perceived Benefits and Purchase Intention. A significant, negative effect was found between Perceived Risks and Purchase Intention, similar to Jarvenpaa et al. (1999), Van der Heijden & Verhagen (2004), Park et al. (2005) and Chang & Chen (2008).

H3 was not supported, although the results were significant, online apparel retail had higher Perceived Risks than groceries. Therefore, Online Product Category had a positive effect on Perceived Risks. The assumption made was the due to groceries lack of online popularity, and apparel’s success, that groceries would have higher Perceived Risks, which in turn would therefore have a negative effect on Purchase Intention. However, this is not the case in this study. This finding could imply that apparel does in fact have a high level of Perceived Risks, similar to Aghekyan-Simonian et al.’s (2012) findings that online apparel is still ridden with perceived risk. This may imply that Perceived Risks are possibly mediated or moderated by other, more dominant factors. Table 8 indicates the t-tests performed on the individual items of variable Perceived Risks. There was a significant difference between means of PR6_SUM (t = -2.466, p <0.01). Respondents therefore found that not being able to touch or feel the product(s) before purchasing was more

Perceived Benefits Low -0.188 0.109 -0.470 -0.021

High 0.097 0.083 -0.030 0.309

Perceived Risks Low -0.281 0.145 -0.648 -0.062

High -0.133 0.142 -0.482 0.104

Unstand. Boot

Effects BootULCI

Temporal Distance

The moderating effect of temporal distance on perceived benefits and risks of online retail categories : a comparative study