The Unravelling of Apparel: Online Shopping Behaviour
Marjolein Kramer[1], Department of Science and Social Science, University College Utrecht, Netherlands
Abstract
The focus of this research is to obtain a better understanding in the determinants of online shopping for apparel and the interaction between consumers' use of the internet for information search and their choice of channel (i.e. brick-and-mortar stores or the internet) for their final purchase. Questionnaires were distributed to over 212 students and the data was analysed with SEM. Firstly, the model applied for both men and women. No interaction was found between in-store shopping for apparel and e-shopping or internet search. Results showed that the Fishbein's Theory of Reasoned Action Model (1975) helped explaining e-shopping for apparel with a few alterations. Attitudes towards e-shopping were mediated by internet search instead of having a direct relation on e-shopping intention. Furthermore, subjective norm not only had an effect on the intention to shop online, but also turned out to affect attitude towards e-shopping. Above all, the internet search component is a critical concept in the behaviour of online shopping for apparel. Opinions of friends and family turned out to be twice as important for men than for women. Overall, the model explained 67.5% of the variance in e-shopping for apparel for women and 71.5% for men.
Keywords: E-shopping, apparel, theory of reasoned action, structural equation modelling, attitude, internet search, subjective norm
Introduction
The internet has spread quickly in the Netherlands: 90% of Dutch households and individuals had access to the internet in 2009 compared to only 63% in 2002 (Statistics Netherlands, 2009). The Netherlands is thus the leader in the European Union in terms of internet access. Furthermore, three-quarters of internet users in 2009 were online shoppers. In the Netherlands online shopping has become an established custom and it is among Europe's leaders in terms of the share of companies that sell products electronically (Statistics Netherlands, 2009). Research carried out by Ernst & Young (2011) found that 52% of Dutch internet users expect to increase further their amount of online shopping. Thuiswinkel.org and Blauw Research (2011) reported that the average annual amount spent online per Dutch person increased from 318 euro in 2002 to 888 euro in 2010. Those numbers have been rising ever since the internet and home computers were introduced.
In this article e-shopping refers to the business-to-consumer (B2C) segment of online purchasing. This way person-to-person websites such as Marktplaats and Ebay will be excluded. With around 8.8 million e-shoppers and an annual B2C turnover of 6 billion euro in 2009, it is becoming more and more important for companies, marketers and managers to obtain a better understanding in the reason behind this online purchasing behaviour (Thuiswinkelmarktmonitor, 2009), particularly given that these numbers are expected to rise even more in the future. The Forrester Research Online Retail Forecast (2011) expects online sales in Western Europe to increase at a 10% compound annual growth rate over the next 5 years.
Moreover, it is important for governments to understand the details and predictors of e-shopping in order to anticipate to this new phenomenon and adapt regulations or laws where necessary. The Distance Selling Act, for example, refers to the sale of goods or services where there is no face-to-face contact between the consumer and seller, as in the case for e-shopping (in effect since 1 February 2001). The most important clauses state the required information that a seller must provide in a transaction and that a consumer has seven days, from the delivery of goods, to return them without consequences (The Economist, 2006). Besides trust, privacy is also an essential issue around e-shopping. The Dutch government, therefore, adapted the Law for Protecting Personal Information to take account of commercial electronic information gathering (The Economist, 2006).
Several research studies acknowledge that consumers are forming different attitudes toward shopping for diverse product categories sold online (Hyllegard et al., 2000). Not only the attitudes, but also the purchase frequency, the choice of shopping mode (e-shopping or in-store shopping) and the physical shopping place vary across product categories (Lenz et al., 2003). This study will therefore focus on e-shopping for apparel, as the number of internet consumers buying apparel online experienced the highest increase between 2005 and 2007 (Nielsen, 2008). Furthermore, according to Thuiswinkel Markt Monitor (2011) fashion and apparel purchases represent a significant portion of e-shopping and thus represent one of the largest emerging e-shopping categories.
Due to the increasing popularity of e-shopping and its long-term potential in the retail industry, many marketers and managers are interested in the change that this will bring about. Most previous research studies about e-shopping focused solely on online buying (intention) (van der Heijden et al., 2003). However, for companies to make optimal use of these outcomes and promote e-shopping, more information is needed. A better understanding of the relationship between consumers' use of the internet for information search and their choice of channel (i.e. brick-and-mortar stores or the internet) for their final purchase is required (Jones and Kim, 2010; Shim et al., 2001: 5). When investigating the connections between e-shopping and traditional in-store shopping it is ever more important to differentiate product types, since the characteristics of products greatly determine the degree to which they are suitable for marketing online and hence their potential for any substitution or complementation effect (Peterson et al., 1997); an additional reason for investigating apparel exclusively. A very important aspect of apparel as an online product category is the fact that apparel purchasing decisions take place at the point of purchase. The dimensions of apparel usually cause consumers to first want to try it on and to touch the fabric.
To obtain a better understanding of the e-shopping phenomenon, not only is its interaction with online search for apparel and in-store shopping valuable, but also what other variables might influence the decision to buy apparel online. The theory of reasoned action (TORA) by Fishbein and Ajzen (1975) is widely used to explain the consumer decision-making process. They postulated that behavioural intention is the function of two components: attitude toward a behaviour and subjective norm. These two components are immediate determinants of intention to perform a behaviour, which is a precursor to behaviour. A later model by Fishbein and Ajzen, the theory of planned behaviour (1985), adds perceived behavioural control as a third determinant of intention to the TORA model. This is defined as an individual's confidence that he or she is capable of performing the behaviour (Ajzen, 1991: 184). However, Kwong and Park (2008: 1476) found in a student sample that the effect of perceived behavioural control on intention was insignificant. They argued that computer literacy and knowledge of internet services are now common skills. Therefore, the TORA model is preferred. Mowen and Minor (1998) explicitly recommended the TORA model for the assessment of purchase intention for high-involvement products such as apparel. The aim of this study is to describe how attitude, subjective norm, online search activities and in-store shopping influence online shopping for apparel. Furthermore, the interaction between in-store shopping and e-shopping is examined.
This study differs considerably from previous research. First, this study investigates the interactions between internet search, in-store shopping, e-shopping intention and actual e-shopping simultaneously. Second, this study is one of the first to test such a model that includes a multiple group analysis to be able to observe gender differences. Third, most studies have been carried out in the United States. Therefore, policymakers and marketers in Europe need further evidence from their own continent. The retail formation in the Netherlands, for example, has almost no large-scale hypermarkets or shopping malls (Farag et al., 2007: 129). Furthermore, 48% of the shopping trips in the Netherlands are made on foot or by bicycle (Ministry of Transport, Public Works, and Water Management, 2004). These specific Dutch retail characteristics might influence the interactions between the variables in this model.
Literature Review
The theory of reasoned action is based on the assumption that human beings are usually quite rational and make systematic use of the information available to them (Fishbein and Ajzen, 1975). The TORA will be used to explore the determinants of e-shopping. As mentioned in the introduction, behavioural intention is a function of two primary determinants: (a) attitude toward e-shopping for apparel, and (b) an individual's perception of normative social pressure to purchase apparel online. Firstly, attitude toward e-shopping is recognised as the positive or negative evaluation of e-shopping for apparel. This is composed of both beliefs that e-shopping for apparel has certain attributes, and a person's evaluation of those beliefs. This definition implies that attitudes develop over time as people gain experience with the behaviour or receive knowledge about the object from other sources (Hasan, 2010: 598). Secondly, subjective norm consists of normative belief - an individual's perception of whether people that are important to them believe they should purchase apparel online - and social approval - an individual's perception of whether these significant people approve or disapprove of e-shopping for apparel. Although apparel might be bought online by customers alone rather than with family or friends, their opinions about e-shopping for apparel will still influence individuals' intentions toward e-shopping (Jones and Vijayasarathy, 1998). Behavioural intention in turn is obtained by the attitude and social norm components and is considered as predictor of actual e-shopping (Kim et al., 2003: 33). Behavioural intention measures how hard people are willing to try, and how much of an effort they are planning to exert in order to shop online for apparel (Ajzen, 1991: 181). Wong et al. (2005) also reported a strong positive relationship between intention and actual e-shopping. Previous studies have found convincing evidence for the TORA model for e-shopping in general. Even the few studies that specifically tested this model for apparel e-shopping have shown that the TORA model holds (Yoh et al., 2003; Kim et al., 2003; Shim and Drake, 1990).
Shim et al. (2001) included an interaction model of consumer information search before the actual purchase by Klein (1998). Klein's interaction model focuses on the essential role of information search for e-shopping in the context of goods that differ based on the type of information sought prior to purchase. In addition, the model expects a person's attitude towards e-shopping to affect the degree of online search. Based on the above review, the hypotheses are summarised as follows:
- H1: Attitudes toward e-shopping for apparel positively influence online purchase intentions for apparel.
- H2: Subjective norm positively influences online purchase intentions for apparel.
- H3: E-shopping intentions positively influence actual e-shopping for apparel.
- H4: Attitudes toward e-shopping for apparel positively influence online search for apparel.
It is remarkable that internet searching has received little attention, as this is seen as an important phase in the shopping process (Farag et al., 2005). An online search is defined as a focused information search via the internet during which a person actively seeks information about apparel; browsing online to acquire initial ideas as well as visiting specific websites in order to compare apparel and/or prices are included (Farag et al., 2005). Since information and preferences are highly related, understanding the consumer information search process is a key element in understanding consumer decision behaviour (Watchravesringkan and Shim, 2003: 2). Watchraversringkan and Shim (2003: 5) found that the stronger the consumers' intentions to search for apparel information online, the more likely consumers were to buy apparel online. Furthermore, Bellman et al. (1999) found that searching online positively affects e-shopping intention. Although no previous research has simultaneously taken internet search, e-shopping intention and actual e-shopping into account, it seems reasonable to assume that internet search, just like attitude and subjective norm, influences e-shopping intention, which in turn affects actual e-shopping for apparel. Consequently, the hypothesis is stated as follows:
- H5: internet search for apparel positively affects e-shopping intention.
E-shopping could substitute, modify or generate traditional in-store shopping (Farag et al., 2007: 140). The substitution of in-store shopping occurs when e-shopping replaces an in-store purchase. Modification refers to in-store shopping that is likely to be altered; that is, the destination choice, mode choice, or timing of the purchase is adjusted because of e-shopping. The generation of trip occurs when e-shopping leads to purchases in-store that otherwise would not have been made. Theoretically, no interaction between in-store and online shopping is also an option. Classifying might be difficult as the interactions could also occur simultaneously (Mokhtarian, 2004). Lenz et al. (2003) suggested that e-shopping will probably lead to a decline in the frequency of in-store shopping trips. This result is supported by Weltevreden and van Rietbergen (2006). A different conclusion was drawn by Mokhtarian (2004) who argues that e-shopping complements in-store shopping. On the other hand, Douma et al. (2004), found evidence for a modification effect; either by searching online before going in-store shopping, or by using the internet to make their in-store trip more efficient. Contradicting results have thus been found concerning the interaction between online and in-store shopping. These mixed results are presumably caused by the different research contexts and settings. The few studies that have been conducted either do not separate online buying from online searching (for example, Ferrell, 2004; Casas et al., 2001), or are only of a descriptive nature and mix product categories that tend to yield vague or inconsistent results (Cao, 2005; Mokhtarian, 2005). Furthermore, previous studies are not consistent with their dependent variable; both e-shopping intention and actual e-shopping have been used. Based upon the previous literature, the hypotheses are stated as follows:
- H6: In-store shopping will have an effect on e-shopping intention for apparel.
- H7: In-store shopping for apparel and actual e-shopping for apparel are correlated.
The majority of internet users (88%) search the internet for product information (BCG, 2001). However, 75% of this group eventually decides to purchase the product in-store (BCG, 2001). Similarly, Ward and Morganosky (2002) indicate that internet search has a positive effect on in-store shopping. However, the contrary of searching for product information in-store before purchasing the product online is not found (Farag et al., 2006a). Consequently, the last hypothesis becomes:
- H8: internet search for apparel positively affects in-store shopping for apparel.
All hypothesised relationship can be seen below in Figure 1.
Figure 1: Structural model including hypotheses
Note that the correlation between in-store shopping and actual e-shopping is obtained by estimating the correlation between the disturbance (d1 and d2) of those endogenous factors. Otherwise this model would not be identified enough for the software to run this model. However, this is equivalent to correlating the two behaviours and will not affect the outcome.
Since it is generally acknowledged that gender has a profound influence upon responses to marketing strategies, specifying the impacts of gender can guide marketers in designing different strategies for different consumers. Various empirical studies have reported that females are generally more interested in and shop more often for apparel than males (Beaudry, 1999; Chiger, 2001; Flynn et al., 2000). However, very few academic studies have focused on gender differences in online shopping. One of the few studies about internet searching show that women are more likely than men to search online for information regarding apparel products (Watchravesringkan and Shim, 2003: 4). On the other hand, men search more often than women when it comes to commercial products and services online (Statistics Netherlands, 2005). Farag et al. (2006a) also examined the duration of shopping and found that for non-daily goods, such as apparel, women shop longer than men. Dittmar et al. (2004: 441) found a more positive attitude towards e-shopping for male than for female. Moreover, Cyr and Bonanni (2005) report that more time and money is spend on online buying by men than by women. Although not all relations between the variables are expected to vary across gender, a multiple group analysis is conducted to examine the gender differences in the model.
Research Methodology
The study had a descriptive, cross-sectional research design. The purpose of this study was to obtain a better understanding in the determinants for apparel e-shopping behaviour. A self-administrative questionnaire was adopted for data collection and a non-probability convenience sample was used consisting of students at the University of Utrecht in the Netherlands. The questionnaires were collected immediately after the students had completed them.
Although the use of students is often criticised due to its higher-than-average proportion of younger adults, the use of a student sample also has some advantages for a study concerning e-shopping, i.e. they will tend to be harbingers of future adoption patterns in the population (Cao and Mokhtarian, 2005). In addition, the OECD (1999) reported that online consumers tend to be young and well-educated compared to traditional consumers. Further, the familiarity of current students with the internet combined with their emerging market power and the likelihood of building customer loyalty provide strong motivations to explore specifically this consumer group (Xu and Paulins, 2005). Ahuja et al. (2003) found similar patterns of findings in online shopping behaviour among students and non-students.
Participation in the study was voluntary and no credit was given in exchange for participation. Students were assured that the survey was anonymous and individual responses could not be identified. It was also made clear to participants that the aggregates of their responses would be used for data analysis purposes only. Finally, participants were assured that neither their participation in the study nor their responses would influence their performance in the course (Hassan, 2010).
Measures
Each student was given a self-administrative questionnaire containing basic demographic information, including age, gender, field of study, nationality and whether they had access to the internet. In addition, three general questions were asked about the relationship between in-store and online shopping or searching. The rest of the questionnaire focused on items that could be used to construct the latent variables. These were generated from previous research papers and were modified to fit the context of e-shopping for apparel where necessary. The questionnaire was originally written in English and then translated into Dutch. See appendix 1 for the entire English version of the questionnaire.
To measure attitude towards e-shopping for apparel thirteen apparel purchasing attributes were identified based on the related literature review (Xu and Paulins, 2005; Watchravesringkan and Shim, 2003; Kim et al., 2003; Crawford, 2000): more merchandise options, higher payment security, more fashionable clothing, more convenience, more time saving, lower prices, better quality of apparel, better customer service, better return policy, better personal advice, more social interaction, more possibilities to compare apparel and sufficient product information. Applied to the present study, attitude towards online purchasing is considered to be a function of the consumer's beliefs about the attributes and the degree of subjective importance a consumer attaches to those attributes (Fishbein and Ajzen, 1975); the expectancy-value model. The beliefs were measured by asking students to indicate how likely it was that e-shopping compared to shopping in-store for apparel would lead to certain attributes. This was measured on a 7-point Likert scale with answers ranging from 'very unlikely' to 'very likely'. The evaluation of their beliefs was measured by asking students to indicate the importance of each attribute when shopping for apparel, irrespective of purchasing online or in-store. Again a 7-point Likert scale was used from 'totally unimportant' to 'very important'. Subsequently, the total score for each attribute was computed by multiplying the two responses to obtain thirteen scores for attitude. These in turn were used in the measurement model in terms of observed indicators to obtain the latent variable attitude.
Subjective norm was constructed of four items asking the student if their friends and family thought they should shop online for apparel; if they would approve e-shopping for apparel; if they would recommend them to buy apparel online; and if their friends and family themselves bought apparel online. All four questions were measured on a 7-point Likert scale from 'not true' to 'true'.
Two questions were asked to measure the amount of online search for each participant. First, as used by Watchravesringkan and Shim (2003), students were asked how likely it was that they would seek information about apparel via the internet rather than from stores, regardless of where they eventually buy the apparel. Answers ranged from 'search entirely by store' to 'search entirely by the internet' on a 7-point Likert scale. The second question was adapted from Farag et al. (2005), asking students how often they searched for information about apparel via the internet. Possible responses were: never, once a year, less than once a month, once a month, several times a month, once a week or several times a week.
The four items that are used to obtain e-shopping intentions are modified from Taylor and Todd (1995a, b). Again a 7-point Likert scale from 'not true' to 'true' was used for the following statements: I intend to use the internet to buy apparel; I plan to use the internet to purchase apparel within the next few months; overall, I would use the internet to buy apparel I need; and, buying apparel via the internet is something I would do.
In assessing the actual e-shopping and in-store purchasing behaviour for apparel two questions per construct were used. Participants were asked how often they purchased apparel online and how often they purchased apparel in-store. Then they were asked how many times they have bought apparel online and how many times in-store since January 2011.
Data Analysis
The data analysis was conducted in a structural equation modelling (SEM) framework with AMOS 18 software, whereby a Maximum Likelihood (ML) method was used. By using SEM an explanation can be given rather than a description of consumers' behaviour and the relationship between online searching, online buying and in-store buying. SEM allows for multiple simultaneous directions of causality, and distinguishes the direct effect and the indirect effect as well as the total effect of an explanatory variable on each dependent variable (Cao and Mokhtarian, 2005).
The measurement model assessed how the latent variables (i.e. attitude, subjective norm, internet search, e-shopping intention, in-store shopping and actual e-shopping) are measured in terms of observed indicators, obtained from the questionnaire. A confirmatory factor analysis (CFA) was conducted for both genders in AMOS 18. This framework provides a means to test the construct validity of item sets to see if they are indirect measures of the hypothesised latent variables (Bollen, 1989). Furthermore, CFA can test whether evidence of construct validity is invariant across males and females and whether characteristics specific to gender that are not related to the construct of interest will influence gender differences (Gregorich, 2006). In order to obtain a scale that can be interpreted for a latent variable, one of the observed indicators is fixed to one.
The structural model applied the causal relationship among these latent variables to test the hypotheses (Kim et al., 2003). Data screening was performed using SPSS 16.0.1 for Windows. Data was checked for normality, linearity, absence of outliers and absence of multi-collinearity. Furthermore, a missing value analysis was conducted in SPSS.
Model Fit
How well the model eventually matches the data is illustrated by the goodness of fit index. There are three different categories of model fit indices, i.e. absolute fit indices, incremental fit indices and parsimony fit indices. At least one fit index of each category will be used.
Absolute fit indices determine how well an a priori model fits the sample data (McDonald and Ho, 2002) and show which of the models have the best fit. They demonstrate how well the model fits in comparison to no model at all (Jöreskog and Sörbom, 1993). A chi-square test is the most common fit measure, but it is only recommended with moderate samples of 100 to 200 (Kääriäinen et al. 2011, Tabachnick and Fidell, 1996). A statistical significant χ2 indicates that a significant amount of observed covariance between items remains unexplained by the model (Cole, 1987). With over 200 cases and many correlations, alternative measures of fit should be taken into account (Hu and Bentler, 1999; Byrne, 2004; Kline, 2005; and Meyers et al., 2006). Therefore, as recommended by Thompson (2004), Kline (2005) and Byrne (2001), the RMSEA was also taken into account, which is the second most reported fit statistic. AMOS calculates confidence intervals of 90% for the RMSEA. Hereby, should the lower bound of a well-fitting model be around zero and the upper limit should be below .07 (Hooper et al., 2008).
Secondly, incremental fit indices are based on the comparison of the fit of a substantive model to that of a null model, which yields unconstrained estimates of the variance of the observed variables (McDonald and Ho, 2002). The fit index mostly used in this category is the Comparative fit index (CFI). It is commonly accepted that a CFI greater than .90 suggests an adequate fit of the model (Gefen et al., 2000; Hair et al., 1998). Besides the CFI, the Tucker-Lewis Index (TLI) will also be used. This index will adjust for complexity of the model. Typically, a TLI value of at least .90 is required to accept the model.
Lastly, the Akaike Information Criterion (AIC) was used from the parsimony fit indices. This index adjusts for sample sizes and takes the model fit as well as the complexity of the model into account. The AIC is not normed to a 0-1 scale and has no cut-off value. Instead the model that generates the lowest value is the most superior. In conclusion, the relative chi-square, RMSEA with its confidence interval, CFI, TLI and AIC will be used to assess the model fit.
Results
Sample Description
Two-thirds of the data was collected in a classroom setting from a sample of college students who were enrolled in several courses at the University of Utrecht in the Netherlands. Furthermore, several students from University College Utrecht in different disciplines were asked to fill out the questionnaire. The remaining one-third of the questionnaires was collected in the library of the University of Utrecht. The sample consisted of 85 (40.1%) men and 127 (59.9%) women. Their ages ranged from 17 to 29, with a mean of 22 years and a standard deviation of 2.0 years. All participants had internet access and used this at least one hour per week. The majority of the participants (44.5%) used the internet between 10 and 20 hours a week and another 36.0% spent more than 20 hours online. Approximately two-thirds of the participants has purchased apparel online in the past. As expected, due to the focus on apparel, more female students than male students have done this before; 56.1% of the males had purchased apparel online before, compared to 70.9% of the female students. The largest number of students, almost one-third, were studying law (31.9%), followed by economic students (22.9%) and students from University College Utrecht, a Liberal Arts and Science college (19.5%). The remaining 25.7% studied in a variety of fields including medicine, theatre, computer science and history. The total sample consisted of 17 different nationalities. Nevertheless, the vast majority were of Dutch nationality (86.8%), far more than the second largest group, Germans (2.8%). Little's chi-square statistic, obtained from the missing value analysis in SPSS, was insignificant (p=0.53), implying that the missing values are missing completely at random. None of the variables had more than 1.5% missing values. Skewness and kurtosis results showed that none of the items were above the recommended cut-off points of |3.0|, implying that there was no univariate non-normality (Kline, 2005).
More than half of the students (58.8%) said they had never searched in-store for apparel and eventually bought this online; 21.7% rarely did this; and only 14.6% of the students answered this question with 'sometimes'. The other way around though, where consumers search for apparel online and eventually purchase this in-store, was used more often. Over one-third (38.7%) stated they sometimes do this and 12.8% of the students do this often or even very often. One interesting difference between male and female worth mentioning is the amount they search online. To the question how often they search for information about apparel online the average answer for men is less than once a month, whereas women do this on average once a month. No significant differences in answers were found when examining the descriptives for different nationalities or fields of study.
Confirmatory Factor Analysis
Principal component factor analysis with varimax rotation was conducted to determine the underlying dimensions of the group of attitude items. Only the factors with an eigenvalue of higher than 1.0 were retained and were used as items to construct the latent variable 'attitude towards e-shopping for apparel'. Since all items loaded above 0.5 on one of the factors, all items were retained (Kim et al., 2003). Factor analysis produced two factors of attitudes toward e-shopping for apparel that accounted for 53.7% of the cumulative variation in attitude, with factor loadings ranging from 0.56 to 0.88 as can be seen in Table 1. The use of two factors is also supported by the scree plot check. Factor 1 was named 'product and convenience' and was composed of seven items. Factor 2 was named 'service' and was composed of the remaining six items. A score for each participant for each attitude factor was obtained by summing raw scores of all items in that factor divided by the number of items in that factor. Cronbach's alpha scores assessing internal consistency of measures were above 0.80, indicating good reliabilities of measures (Yoh et al., 2003). In addition, the KMO statistic of 0.86 showed that the factor analysis yields distinct and reliable factors.
| Factor label and items | Factor loading | Eigen- value | Variance | Cronbach's Alpha |
|---|---|---|---|---|
| Product and Convenience | 5.21 | 28.01 | 0.82 | |
| Merchandise options | 0.67 | |||
| Fashionable clothing | 0.60 | |||
| Convenience | 0.70 | |||
| Time saving | 0.75 | |||
| Lower Prices | 0.61 | |||
| Possibilities to compare apparel | 0.68 | |||
| Sufficient product information | 0.67 | |||
| Service | 1.77 | 25.66 | 0.82 | |
| Higher payment security | 0.52 | |||
| Quality of apparel | 0.61 | |||
| Customer service | 0.81 | |||
| Return policy | 0.56 | |||
| Personal advice | 0.88 | |||
| Social interaction | 0.82 |
Table 1: Factor analysis: Attitude toward e-shopping for apparel
The mean of belief scores for the factor 'product and convenience' (M=4.5) was considerably higher than that for the factor 'service' (M=2.5). Within the 'product and convenience' attitude factor, participants believed that e-shopping for apparel compared to in-store shopping mostly lead to 'time saving' (M=5.4), followed by 'more merchandise options' (M=4.9). The most important attribute was 'lower prices' (M=5.3), followed by 'more possibilities to compare apparel' (M=5.0). In terms of the 'service' attitude factor, the attribute believed most strongly by respondents was 'a better return policy' (M=3.2), followed by 'higher payment security' (M=3.0). The most important attribute was 'better quality of apparel' (M=5.6), followed by 'better return policy' (M=4.6). Although social interaction has the lowest belief score (M=1.6), this item is also seen as the least important attribute when shopping for apparel (M=3.6).
Overall, the scores for attitudes toward products and convenience (M=22.0) were twice as high as attitudes toward service (M=11.5). The attitude towards 'more merchandise options' was the highest (M=24.0), followed by 'more time saving' (M=23.7). Not surprisingly, both of these items fall under the products and convenience factor. Within the service factor 'better quality' received the highest attitude (M=15.8).
Measurement Model
Before turning to the structural model that examines the proposed hypotheses, a more detailed assessment of the measurement model is necessary. All factor loadings were highly significant, p<.001, except for the item 'How often have you purchased apparel in-store since January 2011', as seen in Table 2. Therefore, this item was removed and the remaining item 'How often do you purchase apparel in-store?' was used to replace the latent variable 'In-store shopping'. The latent variable 'actual e-shopping' was initially built up in the same structure. To keep this structure for e-shopping the same as for in-store shopping, the latent variable was replaced by the observed variable how often they bought apparel online.
| Path | B | S.E. | p-value | Beta |
|---|---|---|---|---|
| Male | 39.74 | 147.03 | 0.79 | 2.83 |
| Female | 15.48 | 34.73 | 0.66 | 1.85 |
Table 2: Insignificant indicator for the latent variable in-store shopping.
Subsequently, equality constraints on the measurement model were added to ensure that the same construct was measured in both groups. First, the factor loadings are constrained to be equal across groups. This way metric invariance is tested, which is necessary to ensure that different groups respond to the items in the same way so that the obtained ratings from different groups can be obtained in a meaningful way (Hair et al., 2006; Steenkamp and Baumgartner, 1998); observed item differences then indicate group differences in the underlying latent construct. Second, scalar invariance is examined, which is necessary to make mean comparisons in latent constructs across gender (Meredith, 1993; Teo et al., 2009). To test for scalar invariance the intercepts are equalised across gender. The third and fourth step is to equalise measurement error variances and factor means, respectively, although these are not strictly necessary to conclude invariance across gender.
Because the metric invariance model (Nr 2.) is nested within the unconstrained model (Nr 1.), a χ2 difference test was performed. The model comparison showed that equalising the path coefficients did not make the model significantly worse. The χ2 difference was statistically insignificant at p=0.83, with Δχ2=4.31 and eight degrees of freedom difference. Although the χ2 difference test is widely used to compare the fit of nested models, it is recommended that several other model fit indices are used. Thus, CFI, TLI, RMSEA and AIC were also considered. They all slightly improved and the AIC decreased, meaning that the model fit improved and the complexity decreased (Table 3).
| Nr | Model | df | CFI | TLI | RMSEAa | AIC |
|---|---|---|---|---|---|---|
| 1 | Unconstrained | 96 | .95 | .92 | .06 [.04-.07] | 327.18 |
| 2 | Constrained path coefficients | 104 | .95 | .93 | .05 [.04-.07] | 315.49 |
| 3 | Constrained path coefficients; and intercepts | 116 | .93 | .90 | .06 [.05-.07] | 332.51 |
| 4 | Constrained path coefficients; and intercepts except for one search item | 115 | .95 | .93 | .05 [.04-.07] | 309.46 |
| 5 | Constrained path coefficients; all intercept but one; and error variances | 127 | .92 | .90 | .05 [.05-.07] | 332.82 |
Table 3: Measurement (Factor) Model (The terms between brackets represent the 90% confidence interval for the RMSEA)
With the support of the metric invariance model (nr2.), scalar invariance was tested by constraining the intercepts to be the same across gender. The χ2 difference test was statistically significant, p=.00, with Δχ2(12)=41.02. Thus, the model got significantly worse when equalising the intercepts (see Table 3). The intercept for online search 'How often do you search online' turned out to be responsible for this. Equalising all intercepts except for this search item led to an insignificant χ2 difference test, p=.14 with Δχ2(11)=15.97; also the model fit indices improved. Bryne et al. (1989) and Steenkamp and Baumgartner (1998) have stated that full metric and scalar invariance is not necessary for further test of invariance and substantive analysis, such as comparisons of factor means, to be meaningful. Then the error variances were equalised (except for the search item with an unconstrained intercept). However, this did not improve the model and made it significantly worse, i.e. a χ2 difference test of p=.00 with Δχ2(12)=47.36. Model number 4 in Table 3, with full metric and partial scalar invariance, was thus chosen as final measurement model.
Structural Model
The overall fit indices for the hypothesised model revealed a χ2(77)=279.14 (p<.001). As already mentioned due to its sensitivity other model fit indices must be considered. The χ2/df was 1.73, TLI=0.90, CFI=0.92 and the RMSEA=0.06 with a confidence interval between 0.05 and 0.07. In addition, several regression coefficients were insignificant. The most insignificant path was set to zero, after which the model fit indices were examined. This procedure was continued until no insignificant paths were left. The correlation between the disturbance of in-store shopping and e-shopping for men was deleted first, followed by the path from online search to in-store shopping for men and the path from attitude to e-shopping intention for women. Then the correlation between the disturbances of in-store shopping and e-shopping was also deleted for women, followed by the path from in-store shopping to e-shopping intention for women as well as for men. Subsequently, the remaining insignificant paths - from attitude to e-shopping intention for men and from search to in-store shopping for women - were removed. The model fit indices improved after each deletion and the model did not get significantly worse, except after the last removal of the path from internet search to in-store shopping for women. The corresponding model fit indices can be found in Table 4. Although all the numbers in this paper will generally be rounded off to two decimals, Tables 4 and 5 will be an exception. This is done to obtain a more accurate picture of the change in model fit indices. From Table 4 it can be concluded that the last model, nr. 9, has the best model fit. It must be noted that the chi-square remained significant during all model modifications. Although this was already expected due to the sample size, this must be taken into account.
Subsequently, the remaining paths that were not set to zero for both men and women were equalised. However, since equalising the path between subjective norm and e-shopping intention deteriorated the model fit this path remained free to vary across genders. The model did not get significantly worse with a χ2 difference test of p=0.93 and Δχ2(3)=0.44, and the model fit indices improved (Table 5). The last step was to check the Lagrangian multiplier test with the modification indices to see if the model could be improved by adding a path. The modification indices showed that an improvement in the model was possible by adding a path between subjective norm and attitude. Lastly, this new included path was equalised across genders, which again increased the model fit. All steps are written down in Table 5 where the improvements in model fit indices can be found. For the AIC to be interpreted the model comparisons must be nested. Therefore, the whole procedure was repeated, where the path between subjective norm and attitude was already added and set to zero. Now the AIC could also be used as model fit index.
| Nr | Model | df | Χ2/df | CFI | TLI | RMSEAa | AIC |
|---|---|---|---|---|---|---|---|
| 1 | Measurement model constrained | 161 | 1.734 | .920 | .895 | .059 [.047-.071] | 443.144 |
| 2 | M; Correlation disturbances: In-store → E-shopping | 162 | 1.723 | .920 | .897 | .059 [.047-.070] | 431.144 |
| 3 | M; Search → In-store shopping | 163 | 1.714 | .921 | .898 | .058 [.046-.070] | 429.302 |
| 4 | F; Attitude → E-shopping intention | 164 | 1.704 | .922 | .900 | .058 [.046-.069] | 427.441 |
| 5 | F; Correlation disturbances: In-store shopping → E-shopping | 165 | 1.696 | .922 | .901 | .058 [.046-.069] | 425.795 |
| 6 | F; In-store shopping → Intention | 166 | 1.688 | .922 | .902 | .057 [.045-.069] | 424.289 |
| 7 | M; In-store shopping → Intention | 167 | 1.678 | .923 | .903 | .057 [.045-.068] | 422.290 |
| 8 | M; Attitude → Intention | 168 | 1.672 | .923 | .904 | .057 [.045-.068] | 420.957 |
| 9 | F; Search → In-store shopping | 169 | 1.673 | .923 | .904 | .057 [.045-.068] | 420.748 |
Table 4: Results of structural equation model I (the terms between brackets represent the 90% confidence interval for the RMSEA)
| Nr | Model | df | Χ2/df | CFI | TLI | RMSEAa | AIC |
|---|---|---|---|---|---|---|---|
| 10 | Equalise structural factor loadings | 172 | 1.646 | .924 | .908 | .055 [.044-.067] | 415.184 |
| 11 | Add: Subjective norm → Attitude | 170 | 1.556 | .936 | .921 | .051 [.039-.063] | 400.547 |
| 12 | Equalise last added path | 171 | 1.548 | .936 | .922 | .051 [.039-.063] | 398.672 |
Table 5: Results of structural equation model II (the terms between brackets represent the 90% confidence interval for the RMSEA)
Tables 4 and 5 show that the last model, nr. 12, has the best model fit indices that all point at a reasonable good model fit. The relative chi-square is below the cut-off value of two (χ2/df=1.55). Further, the CFI=0.94 and TLI=0.92 are above the recommended .90 and the 90% confidence level of the RMSEA, from 0.04 to 0.06, remained entirely below 0.07. Lastly, the AIC value for model nr. 12 is the lowest of all values obtained, implying that this model is the best model when taking the model fit as well as the complexity into account.
Tables 6 and 7 include the unstandardised estimates, the standard error, p-value and standardised estimates for the final measurement en structural model, respectively. Since the path coefficients are equalised across gender these unstandardised values are the same for male and female (except for the path between subjective norm and e-shopping intention).
| Path | B | S.E. | p-value | Beta, men | Beta, women |
|---|---|---|---|---|---|
| Intention → Something I would do… | 1.00 | - | - | .92 | .82 |
| Intention → Plan to… | .98 | .06 | *** | .90 | .86 |
| Intention → Overall, I would… | .88 | .05 | *** | .89 | .83 |
| Intention → Intend to… | .97 | .05 | *** | .91 | .88 |
| Search → I would seek entirely by… | 1.00 | - | - | .79 | .76 |
| Search → I search for information | .99 | .11 | *** | .82 | .70 |
| Subjective norm → Family buys online… | 1.00 | - | - | .58 | .49 |
| Subjective norm → Family approves… | .69 | .16 | *** | .38 | .38 |
| Subjective norm → Family recommends… | 1.54 | .23 | *** | .88 | .85 |
| Subjective norm → Family thinks I should… | 1.17 | .18 | *** | .72 | .60 |
| Attitude → Product and convenience | 1.00 | - | - | .79 | .81 |
| Attitude → Service | .60 | .09 | *** | .59 | .67 |
Table 6: Final Measurement Model, *** < .001
| Path | B | S.E. | p-value | Beta, men | Beta, women |
|---|---|---|---|---|---|
| Subjective Norm → Attitude | 3.04 | .75 | *** | .45 | .38 |
| Subjective Norm → Intention | .83/ .43 | .20/ .19 | ***/.02 | .42 | .21 |
| Attitude → Search | .12 | .02 | *** | .55 | .69 |
| Search → Intention | .92 | .12 | *** | .65 | .60 |
| Intention → E-shopping | .46 | .03 | *** | .85 | .82 |
Table 7: Final Structural Model, ***<.001, male/female
Table 7 shows that the relation between e-shopping intention and actual e-shopping is strongest for both men and women with a standardised coefficient of 0.85 and 0.82, respectively. Before returning to the hypotheses it can be concluded that the model applies to both men and women. Only between social influences of family and friends and the intention to buy apparel online there is a noticeable difference between genders. Apparently, opinions of family and friends weigh twice as strong for men (beta=0.42) than for women (beta=0.21). Of the eight hypotheses, four obtained a significant result. Attitude toward e-shopping turned out to have no significant effect on online purchase intentions for apparel as hypothesised in H1. Therefore, H1 was rejected. However, attitude towards e-shopping for apparel has an indirect effect on e-shopping intentions, through online search (beta=0.36 for male and beta=0.42 for female). The other hypotheses that were based on the TORA model, H2 and H3, were both accepted. Subjective norm had a significant positive effect on e-shopping intentions, which in turn had a significant positive effect on the actual e-shopping behaviour. Subjective norm has an indirect effect on e-shopping for apparel of beta=0.49 for male and beta=0.30 for female. Although attitude does not have a direct significant effect on e-shopping intention it significantly affects online search behaviour, as expected by hypothesis 4. Therefore, H4 was accepted. Hypothesis 5 was also accepted, since online search significantly affects purchasing intention online for apparel. In addition, online search has a significant indirect effect on e-shopping for apparel, i.e. beta=0.55 for male and beta=0.49 for female.
Previous studies have often raised questions about the relation between e-shopping and in-store shopping in their discussion. The few studies that did somehow included in-store shopping in the model often obtained contradicting results. Nevertheless, with some improvements and a different model an attempt was made to obtain a better understanding in the role of in-store shopping. However, no significant results were found for the relation between in-store shopping and e-shopping intentions or actual e-shopping, H6 and H7, respectively. Therefore, H6 and H7 were both rejected. Furthermore, internet search did not have a significant effect in predicting the amount of in-store shopping for apparel, leading to the rejection of H8.
Even though several hypotheses proved to be insignificant the model explains 67.5% of the variance in e-shopping for apparel for women and 71.5% for men. For both men and women e-shopping intention has the largest total standardised effect on actual e-shopping (which in this case is the same as the direct standardised effect), i.e. beta=0.85 for male and beta=0.82 for female. internet search has the second largest effect on actual e-shopping, i.e. beta=0.55 for male and beta=0.49 for female. For women the total standardised effect of attitude on e-shopping exceeds that of subjective norm, i.e. beta=0.34 and 0.30, respectively. On the other hand, for men the opposite is true; subjective norm has a larger effect on e-shopping than attitude towards e-shopping for apparel, i.e. beta=0.49 and 0.31, respectively. The final model including standardised regression coefficients can be found on the next page in Figures 2 and 3 for men and women, respectively. Appendix 2 contains the final model with unstandardised regression coefficients. Note that the model for men and women is almost the same since the unstandardised paths are equalised across gender. The only difference is the path between subjective norm and e-shopping intention.
Figure 2: Final model with standardised regression coefficients - Male
[Click on image for full-size version]
Figure 3: Final model with standardised regression coefficients - Female
[Click on image for full-size version]
Discussion
Females are in general more interested in shopping for apparel and like to spend more money on apparel and accessories than men; for example, 71% of all dollars spent in that category in the U.S. for February 2010 was spent by women (comScore e-Commerce Report, 2010). It was therefore no surprise that significantly more female than male students have purchased apparel online before. The fact that females search more often for information about apparel online confirms the findings of Watchraversringkan and Shim (2003: 4). The results obtained from examining attitude support previous findings of Reibstein (2002) who found that customers state and behave as if price is the most important factor in drawing them to a website.
Results showed that this study supported most of Fishbein and Ajzen's Theory of Reasoned Action (1975) by indicating that subjective norm influences e-shopping intention (hypothesis 2), which in turn affects actual e-shopping behaviour (hypothesis 3). The stronger the perceived support by family and friends for e-shopping of apparel is, the more likely the intention to actually purchase apparel online. A deviation from the TORA model though, was the predicted direct relation between students' attitude and intentions toward e-shopping. Attitude toward e-shopping had no significant direct effect on online purchase intentions for apparel as hypothesised in hypothesis 1. This contradicts earlier findings that made use of the TORA model (Yoh et al., 2003; Kim et al., 2003). Instead this relationship was mediated by online search as hypothesised by hypotheses 4 and 5. Online retailers could thus still benefit from informing consumers about these attributes and improve them where necessary to increase the amount of time they will search online for apparel, which in turn will then increase their intentions toward e-shopping for apparel and eventually their actual e-shopping behaviour. The second deviation from the TORA model is the relation between subjective norm and attitude towards e-shopping for apparel. This added path is not an entirely new phenomenon. Although the two variables are separate constructs in the TORA model by Fishbein and Azjen (1975), recent studies proposed they might be positively related (Pan et al., 2003; Venkatesh et al., 2003). Even before those studies Chang (1998) already found an improvement in model fit when adding a path from subjective norm to attitude to the TORA model during his research on unethical behaviour.
An important purpose of this study was to examine the interaction between in-store shopping for apparel and e-shopping for apparel. However, all hypothesised relationships with in-store shopping were rejected as they turned out to be insignificant for both genders (hypotheses 6 and 7). It must be noted that the variables in-store shopping for apparel and e-shopping for apparel were both measured by one question only: 'How often do you purchase apparel online (in-store)'. Besides the fact that it is only one question, this question is also based on personal knowledge and judgment. Future research might be able to use actual numbers registered by (web) shops themselves to construct these variables to obtain a more accurate picture.
Several implications for online retailers of apparel can be drawn. Above all, the internet search component is a critical concept in the behaviour of online shopping for apparel. This suggests that the retail strategy should emphasise the information that can be found on their website. In addition to internet search, subjective norm and attitudes revealed to be a significant indirect predictor of e-shopping for apparel. Therefore, online retailers should pay close attention to social influences from family, friends and other acquaintances. Online retailers should use their knowledge about the importance of social reinforcement by using the word-of-mouth advertising strategy. To improve the attitude towards e-shopping for apparel retailers should emphasise the benefits of this e-shopping behaviour in contrast to purchasing apparel in-store. The three most important attributes of e-shopping according to the student sample were better quality of apparel, lower prices and more possibilities to compare apparel. Online retailers should thus start assure customers that shopping at their website for apparel will lead to this.
Some limitations of the present study need to be acknowledged. First, the type of apparel might influence attitude and subjective norm. These constructs could differ when measured for functional (e.g. rain coat or sport clothes) or expressive apparel (e.g. social clothes or new fashion), but also whether they are used for themselves or for others. It could thus be useful to observe whether the validity of measures and results hold across distinct apparel categories. Secondly, lack of randomness in the sample limits the ability to generalise the point and interval estimates to a larger population. However, the results are still useful since students belong to a large part of the online consumers, which makes them interesting to online retailers. Nevertheless, other populations should also be examined to confirm and expand the obtained results.
Future research should take several points into account. The expansion of Fishbein and Azjen's TORA model (1975) should be taken seriously since the relation between subjective norm and attitude has shown up in earlier studies. Future research should examine this path for different samples and settings to examine whether this relationship is specific for a student sample or can be generalised to the entire Dutch or European population. Since searching online turned out to have a positive effect on e-shopping intentions, future studies should focus on how to convert people that seek information online for apparel to online buyers. This could help online retailers to increase their understanding in the buying process of the customers, which eventually could be used to boost profits.
The current study only provides a snapshot picture. E-shopping for apparel still remains a fairly new phenomenon and the conclusions drawn in this study might not hold for long due to the rapid developments. Although Fogel and Schneider (2010) already conducted a longitudinal study of three months, a study of at least a year would be more helpful in determining long-term patterns and to assess the changes in consumer perceptions towards e-shopping. Although already a large amount of variation is explained by the current model, further studies could also pay attention to the remaining 30% unexplained variance in e-shopping for apparel. Possibilities are perceived risk, previous e-shopping experience and perceived behavioural control.
Lastly, at this time more and more online retailers provide physical service points where customers can pick up, pay for, and return items they ordered online (Weltevreden et al., 2005). Future research should therefore be more precise in specifying the terms e-shopping and in-store shopping. There is an important difference in buying apparel online at a website that does not have a physical store in contrast with websites that also possess a building or store for operations.
List Of Illustrations
Figure 1: Structural model including hypotheses
Figure 2: Final model with standardised regression coefficients - Male
Figure 3: Final model with standardised regression coefficients - Female
List of Tables
Table 1: Factor analysis: Attitude toward e-shopping for apparel
Table 2: Insignificant indicator for the latent variable in-store shopping
Table 3: Measurement (Factor) Model
Table 4: Results of structural equation model I
Table 5: Results of structural equation model II
Table 6: Final Measurement Model, *** < .001
Table 7: Final Structural Model, ***<.001, male/female
Appendix
Questions used in the Questionnaire
1. Gender Male / Female
2. Age …
3. What is your field of study? …
4. What is your nationality? …
5. Do you have access to the internet? Yes / No
| Never | Less than 1 hour | 1 to 5 hours | 5 to 10 hours | 10 to 20 hours | More than 20 hours |
6. On average, about how many hours a week do you spend using the internet?
7. In the past I have purchased apparel online? Yes / No
8. I search for apparel information online and then buy it in a store.
| Never | Rarely | Sometimes | Often | Very often |
9. I search for apparel in-store and then buy it online.
| Never | Rarely | Sometimes | Often | Very often |
10. I have made a shopping trip due to searching apparel information online that I would not have made otherwise.
| Never | Rarely | Sometimes | Often | Very often |
| Circle your answer | Not true | True | |||||
| 11. Friends and family think I should shop online for apparel. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 12. Friends and family approve e-shopping for apparel. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 13. Friends and family recommend me to buy apparel online. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 14. My friends and family buy apparel online. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 15. I intend to use the internet to buy apparel. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 16. I plan to use the internet to purchase apparel within the next few months. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 17. Overall, I would use the internet to buy apparel I need. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 18. Buying apparel via the internet is something I would do. | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| 19. How often do you purchase apparel online? | Never | Rarely | Sometimes | Often | Very often | ||
| 20. How often do you purchase apparel in-store? | Never | Rarely | Sometimes | Often | Very often | ||
21. Since January 2011 I have bought apparel online.
| Never | 1-2 | 3-4 | 5-6 | 7-8 | 9-10 | More than 10 times |
22. Since January 2011 I have bought apparel in-store.
| Never | 1-2 | 3-4 | 5-6 | 7-8 | 9-10 | More than 10 times |
23. Compared to shopping for apparel in a store, how likely is it that e-shopping for apparel will lead to…
| Circle your answer | Very Unlikely | Very Likely | |||||
| More merchandise options | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Higher payment security | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More fashionable clothing | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More convenience | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More time saving | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Lower prices | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better quality of apparel | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better customer service | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better return policy | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Better personal advice | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More social interaction | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More possibilities to compare apparel | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Sufficient product information | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
24. Irrespective of buying online or in-store, when I decide where to shop for apparel the following attributes are important…
| Circle your answer | Totally Unimportant | Very Important | |||||
| More merchandise options | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Higher payment security | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More fashionable clothing | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More convenience | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More time saving | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Lower prices | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better quality of apparel | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better customer service | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| A better return policy | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Better personal advice | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More social interaction | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| More possibilities to compare apparel | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
| Sufficient product information | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
25. I would seek information about apparel via the internet rather than from stores, regardless of where I eventually buy the apparel.
| Search entirely by store | Search entirely by internet | |||||
| 1 | 2 | 3 | 4 | 5 | 6 | 7 |
26. I search for information about apparel via the internet.
| Never | Once a year | Less than once a month | Once a month | Several times a month | Once a week | Several times a week |
Notes
[1] Marjolein Kramer graduated from University College Utrecht in July 2011 with a degree in Economics and Mathematics and she is currently reading for an MSc in Econometrics at Erasmus University Rotterdam.
References
Ahuja, M., B. Gupta and P. Raman (2003), 'An Empirical Investigation of Online Consumer Purchasing Behaviour', Communications of the ACM, 46 (12), 145-51
Ajzen, I. (1985), 'From intentions to actions: A theory of planned behaviour', in J. Kuhl and J. Beckman (eds.), Action-control: From cognition to behaviour, Heidelberg: Springer, pp. 11-39
Ajzen, I. (1991), 'The theory of planned behaviour', Organizational Behaviour and Human Decision Processes, 50, 179-211
Ajzen, I. (2006), Constructing a TpB Questionnaire: Conceptual and Methodological Considerations, http://www.people.umass.edu/aizen/pdf/tpb.measurement.pdf, accessed 21 February 2012
BCG (Boston Consulting Group) (2001), The multi-channel consumer: the need to integrate online and offline channels in Europe, http://www.bcg.com/publications/files/MultichannelConsumer_summary.pdf, accessed 19 May 2011
Beaudry, L. (1999), 'The consumer catalog shopping survey', Catalog Age, 16 (6), A5-A18
Bellman, S., G. Lohse and E. J. Johnson (1999), 'Consumer Buying Behaviour on the Internet: Finding From Panel Data', Journal of Interactive Marketing, 14, 15-29
Bollen, K. (1989), Structural Equations With Latent Variables, New York: John Wiley and Sons
Bryne, B., R. Shavelson and B. Muthen (1989), 'Testing for the Equivalence of Factorial Covariance and Mean Structures: The Issue of Partial Measurement Invariance', Psychological Bulletin, 105, 456-66
Byrne, B. (2001), Structural equation modeling with AMOS: Basis concepts, applications and programming, New Jersey: Lawrence Erlbaum Associates (originally published by Psychology Press in 2000)
Byrne, B. (2004), Structural Equation Modeling with AMOS: Basic Concepts, Applications, and Programming, Mahwah, NJ: Erlbaum (originally published by Psychology Press in 2000)
Cao, X. M. (2005), The intended and Acutal Adoption of Online Purchasing: A Brief Review of Recent Literature, Davis: Institute of Transportation Studies
Carini, A. (2011), European Online Retail Forecast, 2010 to 2015, http://www.forrester.com/European+Online+Retail+Forecast+2010+To+2015/fulltext/-/E-RES58597, accessed 20 May 2011
Casas, J., J. Zmud and S. Bricka (2001), 'Impact of shopping via internet on travel for shopping purposes', Washington, DC: the 80th Annual Meeting of the Transportation Research Board
Chang, M. (1998), 'Predicting Unethical Behaviour: A Comparison of the Theory of Reasoned Action and the Theory of Planned Behaviour', Journal of Business Ethics, 17, 1825-34
Chiger, S. (2001), 'Consumer Shopping Survey: Part 2', Catalog Age, 2001
Cole, D. (1987), 'Utility of Confirmatory Factor Analysis in Test Validation Research', J Consult Clin Psychol, 55, 584-94
ComScore e-Commerce Report. (2010). U.S. Consumers, Non-travel Internet Purchases, http://www.comscore.com/Press_Events/Press_Releases/2010/2/comScore_Reports_U.S._E-Commerce_Spending_in_Q4_2009_Reached_39_Billion_Up_3_Percent_vs._Year_Ago, accessed 3 July 2011
Crawford, F. (2000), 'Consumer Relevancy: Connecting With the 21st Century Market', Global Online Retailing Report, 40-5
Cyr, D. and C. Bonanni (2005), 'Gender and website design in e-business', International Journal of Electronic Business, 3 (6), 565-82
Dittmar, H., K. Long and R. Meek (2004), 'Buying on the Internet: Gender Differences in On-line and Conventional Buying Motivation', Sex Roles, 50 (5-6), 423-44
Douma, F., K. Wells, T. A. Horan and K. Krizek (2004), ICT andTtravel in the Twin Cities MetropolitanAarea: Enacted Patterns Between Internet Use and Working and Shopping Trips, Washington DC: the Transportation Research Board 83rd Annual Meeting
Ernst and Young (2011), Onderzoekresultaten ICT Barometer over Online Winkelen, http://ict-barometer.nl/_files-cms/File/ICT%20Barometer%20over%20online%20winkelen-mei%202011.pdf, accessed 2 June 2011
Farag, S., T. Schwanen and M. Dijst (2005), 'Online Searching and Buying and Their Relationship With Shopping Trips Empirically Investigated', Transportation Research Record 1926, pp. 242-51
Farag, S., J. Weltevreden, T. van Rietbergen, M. Dijst and F. van Oort (2006a), 'E-Shopping in the Netherlands: Does Geography Matter?', Environment and Planning, 33, 59-74
Farag, S., K. J. Krizek and M. Dijst (2006b), 'E-Shopping and Its Relationship With In-store Shopping: Empirical Evidence From the Netherlands and the U.S.', Transport Reviews, 26, 43-61
Farag, S., T. Schwanen, M. Dijst and J. Faber (2007), 'Shopping Online and/or In-store? A Structural Equation Model of the Relationships Between E-shopping and In-store Shopping', Transportation Research A, 41, 125-41
Ferrell, C. (2004), Home-based Teleshoppers and Shopping Travel: Do Teleshoppers Travel less?, Washington, DC.: 83rd Annual Meeting of the Transportation Research Board
Fishbein, M. and I. Ajzen (1975), Beliefs, Attitude, Intention, and Behaviour: An Introduction to Theory and Research, Reading, MA: Addison-Wesley
Flynn, L., R. Goldsmith and W. Kim (2000), 'A Cross-cultural Validation of Three New Marketing Scales for Fashion Research: Involvement, Opinion Seeking, and Knowledge', Journal of Fashion Marketing & Management, 4 (2), 110-20
Fogel, J. and M. Schneider (2010), 'Understanding Designer Clothing Purchases Over the Internet', Journal of Fashion Marketing and Management, 14(3), 367-96
Gefen, D., D. Straub and M. C. Boudreau (2000), 'Structural Equation Modeling and Regression: Guidelines for Research Practice', Communication of the Association for Information Systems, 4 (7), 1-77
Gregorich, S. (2006), 'Do Self-Report Instruments Allow Meaningful Comparisons Across Diverse Population Groups? Testing Measurement Invariance Using the Confirmatory Factor Analysis Framework', Med. Care, 44 (11 Suppl 3), 78-94
Hair Jr., J., R. Anderson, R. Tatham and W. Black (1998), Multivariate Data Analysis, Upper Saddle River, NJ: Prentice-Hall (originally published by Academic Press Inc in 1979)
Hair, J. J., W. Black, B. J. Babin, R. E. Anderson and R. L. Tatham (2006), Multivariate Data Analysis, New Jersey: Prentice-Hall International (originally published by Academic Press Inc in 1979)
Hasan, B. (2010), 'Exploring Gender Differences in Online Shopping Attitude', Computers in Human Behaviour, 26, 597-601
Hoyle, R. H. (1995), In Structural Equation Modeling: Concepts, Issues and Applications, Thousand Oaks, CA: Sage
Hooper, D., J. Coughlan and M. Mullen (2008), 'Structural Equation Modelling: Guidelines for Determining Model Fit', Electronic Journal of Business Research Methods, 6 (1), 53-60
Hu, L. and P. Bentler (1999), Evaluating Model Fit. In Structural Equation Modeling: Concepts, Issues and Applications, Thousand Oaks, CA: Sage
Hyllegard, K., S. Lokken, M. L. Damhorst, H. N. Lakner and H. Bastow-Shoop (2000), 'E-tailing: United States Rural Consumer's Willingness to Adopt the Internet for Information Search and Purchasing', The Netherlands: The European Institute of Retailing and Services Studies
Jöreskog, K. and D. Sörbom (1993), LISREL 8: Structural Equation Modeling with the SIMPLIS Command Language, Chicago: IL: Scientific Software International Inc
Jones, J. M. and L.R. Vijayasarathy (1998), 'Internet Consumer Catalog Shopping: Finding From an Exploratory Study and Directions for Future Research', International Research: Electronic Networking Applications and Policy, 8 (4), 322-30
Jones, C. and S. Kim (2010), 'Influences of Retail Brand Rrust, Off-line Patronage, Clothing Involvement and Website Quality on Online Apparel Shopping Intention', International Journal of Consumer Studies, 34, 627-37
Kääriäinen, M., O. Kanste, S. Elo, T. Pölkki, J. Miettunen and H. Kyngäs (2011), 'Testing and Verifying Nursing Theory by Confirmatory Factor Analysis', Journal of Advanced Nursing, 67 (5), 1163-72
Kim, Y., E. Kim and S. Kumar (2003), 'Testing the Behavioural Intentions Model of Online Shopping for Clothing', Clothing and Textiles Research Journal, 21 (1), 32-40
Klein, L. R. (1998), 'Evaluating the Potential of Interactive Media Through a New Lens: search versus experience goods', Journal of Business Research, 41, 195-203
Kline, R. B. (2005), Principles and Practice of Structural Equation Modeling, New York: Guilford Press (originally published by Guilford Press in 1998)
Kwong, S. W. and J. Park (2008), 'Digital Music Services: Consumer Intention and Adoption', The Service Industries Journal, 28 (10), 1463-81
Lenz, B., T. Luley and W. Bitzer (2003), 'Will Electronic Commerce Help to Reduce Traffic in Agglomeration Areas?' Transportation Research Board 82nd Annual Meeting
McDonald, R. and M-H. Ho (2002), 'Principles and Practice in Reporting Statistical Equation Analyses', Psychological Methods, 7 (1), 64-82
Meredith, W. (1993), 'Measurement Invariance, Factor Analysis and Factorial Invariance', Psychometrika, 58, 525-43
Meyers, L., G. Gamst and A. Guarino (2006), Applied Multivariate Research. Design and Interpretation, Thousand Oaks, CA: Sage Publications
Ministry of Transport, P. W. (2004), Passenger Transport in the Netherlands: Key Figures 2004, http://www.eurorap.org/library/pdfs/Netherlands_RoadSafety2004.pdf, accessed 25 May 2011
Mokhtarian, P. L. (2004), 'A Conceptual Analysis of the Transportation Impacts of B2C E-commerce', Transportation, 31, 257-84 .
Mowen, J. and M. Minor (1998), Consumer Behaviour, Upper Saddle River, NJ: Prentice-Hall (originally published by Prentice-Hall in 1997)
Nielsen Company (2008), Trends in Online Shopping, A Global Nielsen consumer report
Organisation for Economic Co-operation and Development (1999), The Economic and Social Iimpact of E-commerce, http://www.oecd.org/dataoecd/3/12/1944883.pdf, accessed 29 May 2011
Pan, C., S. Sivo and J. Brophy (2003), 'Students' Attitude in a Web-enhanced Hybrid Course: A Structural Equation Modeling Inquiry', Journal of Educational Media and Library Sciences, 41, 181-94
Peterson, R. A., S. Balasubramanian and J. Bronnenberg (1997), 'Exploring the Implications of the Internet for Consumer Marketing', Journal of Academy of Marketing Science, 25 (4), 329-46
Reibstein., D. (2002), 'What Attracts Customers to Online Stores, and What Keeps Them Coming Back?', Journal of the Academy of Marketing Science, 30 (4), 465-73
Salomon, I. (1986), 'Telecommunications and Travel Relationships: A review', Transportation Research, 20 (A), 223-38
Shim, S. and M. Drake (1990), 'Consumer Intention to Utilize Electronic Shopping', Journal of Direct Marketing, 4 (Summer), 22-33
Shim, S., M. Eastlick, S. Lotz and P. Warrington (2001), 'An Online Pre-purchase Intentions Model: The role of Intention to Search', Journal of Retailing, 77, 397-416
Statistics Netherlands, Voorburg (2009), The Digital Economy 2009 CBS, http://www.cbs.nl/NR/rdonlyres/E4311D6B-6BE6-4996-A4AB-804FC0A07A4C/0/2009p38pub.pdf, accessed 28 May 2011
Steenkamp, J-B. E. and H. Baumgartner (1998), 'Assessing Measurement Invariance in Cross-National Consumer Research', Journal of Consumer Research, 25, 78-90
Tabachnick, B. and L. Fidell (1996), 'Using Multivariate Statistics', New York: Harper Collins
Taylor, S. and P. Todd (1995a), 'Understanding Information Technology Usage: A Test of Competing Models', Information Systems Research, 6 (2), 144-77
Taylor, S. and P. Todd (1995b), 'Decomposition of Cross Effects in the Theory of Planned Behaviour: A Study of Consumer Adoption Intentions', International Journal of Research in Marketing, 12 (2), 137-55
Teo, T., C. Beng Lee, C. Sing Chai and S. Wong (2009), 'Assessing the Intention to Use Technology Among Pre-Service Teachers in Singapore and Malaysia: A Multigroup Invariance Analysis of the Technology Acceptance Model (TAM)', Computer & Education, 53, 1000-09
The Economist (2006), Netherlands: Overview of E-commerce, http://www.ebusinessforum.com/index.asp?layout=rich_story&doc_id=9776&title=Netherlands%3A+Overview+of+e-commerce&categoryid=29&channelid=4, accessed 30 May 2011
Thuiswinkel Markt Monitor (2009), Online Thuiswinkelmarkt Groeit 24% Ondanks Financiële crisis, http://www.thuiswinkel.org/bedrijven/nieuws_publicaties.aspx?subnavid=1&id=13622, accessed 29 May 2011
Thuiswinkel Markt Monitor (2011), Gemiddelde Uitgave per Online Koper, http://www.thuiswinkel.org/gemiddelde-uitgave-per-online-koper, accessed 3 June 2011
Thompson, B. (2004), Exploratory and Confirmatory Factor Analysis: Understanding Concepts and Applications, Washington, DC: American Psychological Association
Van der Heijden, H., T. Verhagen and M. Creemers (2003), 'Understanding Online Purchase Intentions: Contributions from Technology and Trust Perspectives', European Journal of Information Systems, 12, 41-48
Venkatesh, V., M. Morris, G. Davis and F. Davis (2003), 'User Acceptance of Information Technology: Toward a Unified View', MIS Quarterly, 27, 425-78 .
Ward, M., and M. Morganosky (2002), 'Consumer Acquisition of Product Information and Subsequent Purchase Channel Decisions', Advances in applied microeconomics: The economics of the internet and E-Commerce, 231-55
Watchravesringkan, K., and S. Shim (2003), 'Information Search and Shopping Intentions Through internet for Apparel Products', Clothing and Textiles Research Journal, 21 (1), 1-7
Weltevreden, J., and T. van Rietbergen (2006), 'City Centre Shopping in the Internet Era: The Case of the Netherlands', paper submitted for publication in Journal of Economic and Social Geography
Wheaton, B., B. Muthen, D.F. Alwin and G. Summers (1977), 'Assessing Reliability and Stability in Panel Models', Sociological Methodology, 8 (1), 84-136
Wong, T., W. So and D. Sculli (2005), 'Factors Affecting Intentions to Purchase via the Internet', Industrial Management and Data Systems, 105 (9), 1225-44
Xu, Y., and V. Paulins (2005), 'College Students' Attitudes Toward Shopping Online for Apparel products: Exploring a Rural Versus Urban Campus', Journal of Fashion Marketing and Management, 9 (4), 420-33
Yoh, E., M. Damhorst, S. Sapp and R. I. Laczniak (2003), 'Consumer Adoption of the Internet: The Case of Apparel Shopping', Psychology and Marketing, 20 (12), 1095-118
To cite this paper please use the following details: Kramer, M. (2012), 'The unravelling of apparel: online shopping behaviour', Reinvention: a Journal of Undergraduate Research, Volume 5, Issue 1, http://www.warwick.ac.uk/reinventionjournal/issues/volume5issue1/kramer Date accessed [insert date]. If you cite this article or use it in any teaching or other related activities please let us know by e-mailing us at Reinventionjournal@warwick.ac.uk.