The Effect of Internet Trust and Social Influence towards Willingness to Purchase Online in Labuan , Malaysia

This study is an empirical research one-commerce that examines the willingness of consumers to purchase online. The study focuses on two influential components that mediate the relationship between online shoppers and online vendors, namely, trust in the Internet structure and social influence. By using a convenience sample size of 115 from government servants and employees of private sectors residing in the Federal Territory Labuan who are computer literate, have access to Internet, and who possess credit card(s), this study aims to examine the relationship between the two variables of trust in the Internet structure and susceptibility to social influence on their willingness as consumers to purchase online. The findings show that trust in the Internet structure and susceptibility to social influence are significantly related to willingness to purchase online. However, hierarchical linear regression analysis had provided insignificant influence between trust in the Internet structure and willingness to purchase online with social influence as a moderator.


Introduction
In the modern 21 st century era, the development of information technology (IT) has certainly brought about far-reaching effects on our society today.The development of computers of increasing speed and processing capabilities has accelerated economic development while creating new opportunities and sectors of activity amidst an increasingly competitive environment.As stressed by Kotler (1987) and Quelch and Klein (1996), this has also caused companies to transform their production systems, product promotion and relationships with customers, producing almost constant opportunities for innovation.
As a result of this rapid growth of IT, Business to Consumer (B2C) electronic commerce (e-commerce) has become a large and essential segment of the new digital economy, which is essentially the ordering of goods and services by individual consumers and making payments for the orders through the Internet.Zwass (2003) explains e-commerce as the sharing of business information, maintaining business relationships and conducting business transactions through telecommunication networks; whilst Pandya and Dholakia (2005) define B2C e-commerce as the new technology-driven means to promote retail and distribution services to consumers.
In order to successfully enable e-commerce, the factor of trust certainly plays a crucial role here, as buyers and sellers typically do not see each other while transacting online, unlike in the offline world.How does a buyer pay a seller he has never seen in a one-time transaction for something he cannot physically inspect until after it is paid for and delivered?Trust is the key; as stated by Dasgupta (1988), "trust is of much importance precisely because its presence or absence can have a strong bearing on what we choose to do and, in many cases, what we can do." To address this issue of e-commerce risk perception by consumers, significant research was performed by (Kim & Prabhakar, 2000;Salam et al., 2003) to understand and minimize the incidence and consequences of risk.Their studies have inferred that these risks would significantly impact the success and viability of the virtual transaction environment.
One of the major factors that could impede the growth of e-commerce is the concern for information privacy.(FTC, 2000;Hoffman et al., 1999;Malhotra et al., 2004).Although many online businesses have broadly implemented online privacy policies designed to signal trustworthy organizational practices (Meinert et al., 2006;Pollach, 2006) and there has also been evidence that these policies do positively influence users' confidence in buying online (Earp & Baumer, 2003); many users still believe that providing information online can result in personal privacy problems (Sinclaire et al., 2006).
The second major factor impeding the growth of e-commerce would be the fear of information security itself (Hawkins et al., 2000;Roberts, 2004).Security issues continue to present challenges to retaining current online customers and attracting new ones for all online organizations in general, whether for commercial entities or the government, and especially for online banks in particular (Rombel, 2005).Nearly 60 percent of the respondents to a recent Gartner survey said they were concerned or very concerned about their online security (Roberts, 2004).Some examples of the information hazards that have occurred include leaks in personal data, security breaches, software flaws, Internet scam and identity theft, all of which frequently make the headlines today.These dangers would certainly influence Internet users' online behaviour in increasing protection of their privacy and minimizing their exposure to security threats.According to the Consumers Report (2005), it was found that 9 out 10 United States of America's Internet users over 18 have made changes to their online behaviour due to fear of identity theft, whilst 30 percent said they have reduced their overall use of Internet, followed by 25 percent stating they have stopped buying things online, and another 29 percent reporting they have cut back on how often they buy things online.
To counter the argument against information security, some researchers have argued that the reliance on the e-commerce process itself has replaced the perception of trust (Shneiderman, 2000), while Friedman et al. (2000) supported the views that even on the Internet, people trust people, not technology.Furthermore, the e-commerce industry is showing signs of approaching maturity; as pointed out by Crockett (2006), who predicted that the number of households with Internet access will only increase by 3 percent in the next three years, i.e. from 64 percent in 2006 to 67 percent in 2009.
On the other hand, the development of e-commerce could still be in a volatile stage, as a survey by Cole (2003) of American Internet users and non-users found that fewer adult Internet users have purchased online in 2002 than in the previous two years, and that more than half of Internet purchasers spent less than they intended when shopping online.Additionally, the Boston Consulting Group estimated that 65 percent of Internet buyers, accounting for 80 percent of the dollar value of prospective purchases, terminated their transaction before checking out (Maravilla, 2001).This evidence highlights the importance and significance of understanding the behavior of online purchasers in e-commerce.
Thus, this study attempts to examine the possible change of consumers' online purchase behavior arising from their trust in the Internet structure and social influence by using data collected from people in the Federal Territory of Labuan 1 .The aim of this paper is to examine the relationship between trust in the Internet and social influence towards consumer willingness to purchase online.

Literature review
The last thirty years has witnessed a tremendous increase in the strength of e-commerce both from the perspective of researchers as well as real-life practitioners, where unlike before, e-commerce was a relatively unknown term to many.However, the willingness of consumers in providing their personal information or their inclination to purchase online remains the major obstacle to the rapid growth of e-commerce in this informative age.
The pioneer studies on e-commerce were initiated by two models, namely the Elaboration Likelihood Model by Petty and Cacioppo's (1986), and Chaiken and Eagly's (1983) Heuristic-Systematic Model.The objectives of both studies were to identify key factors that influence online information-giving behavior.It was identified that relevant reference groups and media are the primary social influences on consumer willingness to provide information online or to purchase online.In addition, there are six other studies that addressed consumer willingness to complete a transaction on the Internet.These studies had identified the factors of trust in the online store (Jarvenpaa et al., 2000), trust in the vendor (Pennington et al., 2003), trust in organizational practices (Smith et al., 1996) and the web site features (Belanger et al., 2002;Gefen & Straub, 2004;Pennington et al., 2003), which can also be characterized as the consumer's perception of the online firm.The study of user trust characteristics that include risk beliefs was covered by Malhotra et al. (2004), while studies on user attitude and risk perception (Jarvenpaa et al., 2000), and individual trusting character (Gefen et al., 2003) were also based on risk beliefs or tendencies.
Evidence of a link between the positive perception of Internet privacy and security and the desire for making Internet purchases was covered in two studies by George (2002George ( , 2004)).The specific factors of trust in the Internet channel were addressed in a model of consumer trust in Internet shopping (Lee & Turban, 2001) which includes a construct of trust in the computerized medium based on perceived technical competence, perceived system performance and user understanding of the system itself.In a similar model of e-commerce relationship trust, McKnight and Chervany (2001) discovered institution-based structural assurance as a factor affecting trust-related Internet behaviour.In this said research, the structural assurance was characterized as "technological Internet safeguards" such as encryption.The structural assurance aspect of the institution-based trust can be compared to the "technology trust" contract defined by Ratnasingham and Povlou (2002) as the subjective probability by which users believe that the underlying technology infrastructure is capable of facilitating transactions securely according to their expectations.
Generally in IT studies, social influence is referred to in some literature as normative belief structures or social norms, where it is frequently decomposed into relevant reference groups.Taylor and Todd (1995) decomposed the sources of social influence into three groups of people: peers, superiors and subordinates.According to their study, influence from peers and superiors were both found to be significantly related to the subjective norm construct that leads to behavioural intention and usage of technology (Taylor & Todd, 1995).Another study conducted by Lewis, Agarwal, and Sambamurthy (2003) analyzed social influence from organizational peers, informal circles, professional peers, supervisors, and senior leaders, only to find no significant effect from any of these sources of social influence.
However, in the context of e-commerce, Limayem et al. (2000) further identified three other components of social influence namely friends, family, and media, in a longitudinal study that examined online buying intentions.Results from this study indicated that media had the strongest influence on online shopping, followed by family influences.However, the influence of friends had no significant effect on the intention to shop online (Khalifa & Limayem, 2003;Limayem et al., 2000).Hwang (2005) found that all these same three components of social influence (friends, family, media) were significantly related to online trust, while Bhattacherjee (2000) found external influences in the form of mass media such as news reports and popular press to have a huge effect on the subjective norms leading to the acceptance of e-commerce.
Predictors on whether an individual will buy online or otherwise were investigated by Bellman, Lohse and Johnson (1999).They found that the most crucial determinant of buying on the web was based on previous Internet usage behavior, such as using the Internet to search for product information, while Vellido et al. (2000) found that the willingness to purchase online was best predicted by consumer risk perception of shopping on the Internet.Furthermore, Bhatnagar et al. (2000) developed a two-part definition of perceived risk to examine its impact on the willingness to purchase online.The two types of risks were defined as product category risk and financial risk.Their results indicated that increases in both types of risks generally decreased the likelihood of purchasing online.
The actual motives for online shopping have long been the focus of consumer and retail researchers.Babin et al. (1994) identified two dominant shopping motives: shopping for fun (hedonic) and shopping with a goal in mind (utilitarian).Wolfinbarger and Gilly (2001) also agreed on these two motives being typical of online shoppers.Goal-oriented shoppers were found to value four attributes of online shopping, namely (1) convenience, (2) information, (3) selection and (4) the ability to control the shopping experience.Meanwhile, the fun-oriented shoppers reported browsing the Internet for fun through auctions, visits to hobby-related sites and bargain hunting.Internet shopping for fun takes advantage of the Internet as a vast repository of information.As the Internet matures, its importance as a source of product information is a major benefit of online shopping (Evans & Wurster, 1999).Though information-rich, the Internet still does not seem to be used as a total substitute for other sources of information.For example, Ramaswami et al. (2000) reported that online shoppers of financial products used both online and offline channels in their information search activities.Furthermore, in a study of new car buyers, Ratchford et al. (2001) found that heavy users of Internet sources were also heavy users of printed sources of information such as car ratings books and dealer brochures.Bhatnagar et al. (2000) found that customers' perceived convenience of shopping on the Internet had a positive impact on their willingness to purchase online.However, Ramaswami et al. (2000) found that for those who use online information sources for buying financial products, shortage of time availability was not associated with the propensity to conduct an online search for these products or to purchase these products online.In other words, those that were hard-pressed for time did not use online shopping more than those who had more free time available.This same result was supported by another research conducted by Alreck and Settle (2002) who found a difference between consumer perception and behavior towards Internet shopping and time saving behavior.
While the characteristics of the shoppers' perceived risk are essential in predicting the willingness to purchase online, Internet shopping is also impacted by the type of products being sold online.According to Bhatnagar et al. (2000), it was discovered that high-risk products such as products that are technologically complex or high priced were less likely to be purchased online.Vijayswathy (2002) found that the willingness to purchase online was inversely related to the tangibility of the product.High cost intangible items such as auto financing generated a higher intention to purchase rating than high cost tangible products such as household appliances.Rosen and Howard (2000) hypothesized that homogeneous goods such as books and music are more suited to online sales compared to differentiated products that require a physical inspection.

Theoretical frame work and research methodology
Based on the early discussion, researchers of the online trust literature provided evidence that users' belief in the trustworthiness of the online medium is linked to the willingness to provide personal information (Lee & Turban, 2001) and the intention to make online purchases (George, 2004).Research by Gefen and Straub (2004) indicated that the inclusions of features that convey a sense of human presence are positively related to individuals' willingness to provide personal information online.Furthermore, social influence (friends, family, and media) is strongly associated with willingness to provide personal information online (Bhattacherjee, 2000;Hwang, 2005;Limayem et al., 2000) and increased purchasing likelihood (Noteberg et al., 2003) Thus, this study considers the relationship between two main independent variables and consumers' willingness to buy goods or services via Internet.The two independent variables are (1) trust in the Internet and (2) susceptibility to social influence.The dependent variable of this study is willingness to purchase online.This research framework is adopted from the study of Sinclaire ( 2007) "Initial Trust Formation in B2C E-commerce".The research framework is present in figure 1.
Figure 1 illustrates a schematic diagram of the research model to be employed in this study.Definition of the variables used in this study: The dependent variables are trust in the Internet and social influence, while the independent variable is willingness to purchase online.
Trust in the Internet infrastructure is defined as trust in the safety and integrity of the fundamental security measures used to protect personal information during online transactions (McKnight & Chervany, 2001).Social influence is defined as user susceptibility to social or interpersonal influence and the tendency of a person to change their online information-giving behavior as a result of social pressure (McGuire, 1968).Willingness to purchase online refers to self-report measure of willingness to purchase goods or services to satisfy needs and desires (Belch and Belch, 2004) via the Internet.

Research hypotheses
Based on the theoretical framework in Figure 1, five research hypotheses were developed.H1: Trust in the Internet is positively related to willingness to purchase online.H2: Susceptibility to social influence is positively related to willingness to purchase online.H3: Susceptibility to social influence is positively related to trust in the Internet.H4: Trust in the Internet is positively linked to susceptibility social influence.H5: The interaction of trust in the Internet and susceptibility to social influence is positively linked to willingness to purchase online.(Sincethere are evidence of interaction effects, these four hypotheses address possible mediation and/or moderation effects among the four constructs in the research model).

Research approach
This study is based on empirical approach as it attempts to analyze the relationship between the independent variable of willingness to purchase online and the dependent variables of Internet trust and social influence.However, this study will be an exploration study as there are not many studies conducted.Thus, dependent variables will also be the moderating and intervening variables.

Sample of analysis
The population of this study will be using convenience samples among government servants and employees of private sectors residing in Federal Territory Labuan, Malaysia.However, the samples must equate with computer knowledge, have an internet access and holding and using credit cards for shopping.These three elements are prerequisites for e-commerce application to be successful.

Data collection method
The method of data collection was utilizing non-probability-sampling technique, that is, convenience sampling due to time and budget constraints.The questionnaires were distributed to the public personally, through family members and friends.The targeted group was government servants and employees of private-sectors as most of them are computer literate and holding credit cards.

Data analysis
For the data analysis of this study, the Statistical package for Social Sciences (SPSS) version 14.0 will be used firstly to obtain the descriptive statistics such as frequencies, mean, and standard deviations.Subsequently, this will be followed by reliability tests to evaluate the consistency in each item of the variables used.The Cronbach's alpha coefficient is expected to be 0.6 or more to establish goodness or accuracy measures.The hypothesis testing based on simple and multiple regression analysis were conducted to test the relationship between the independent and dependent variables.In addition, to observe the effects of moderating variables on the relationship between the independent and dependent variables, Hierarchal Regression was conducted.

Questionnaire
The questionnaire for this research was adapted from the questionnaire constructed and used by Dr. Jollean K. Sinclaire in the research on Initial Trust Formation in B2C E-commerce with some minor modifications.The questionnaire was presented in dual languages, English and Bahasa Malaysia.The said questionnaire was divided into two sections.Section I was to collect information relating to demographic items.There were 8 questions where the respondents were required to select one answer or complete the blank.Section II was designed to collect information relating to the Trust in the Internet, Social Influence and Willingness to purchase online.This section contained 11 items whereby the respondents were required to choose one answer out of five alternatives.The scale used was called the Likert scale whereby the five alternatives are "strongly disagree, disagree, neutral, agree, and strongly agree.

Overview of data collected
A total of 200 questionnaires were distributed to the respondents.From the total of 200 questionnaires, 115 were returned on time for the analysis process.This represents an average response rate of 57.5 percent.The 115 respondents consist of 69 males and 46 females, aged between 25 to 60 years old, with the mean value of the age of 37.0957.The respondents comprise 48 Malays, 28 Chinese, 2 Indians, 22 Bumiputras and 15 foreigners.The Bumiputras comprise Sabah and Sarawak natives, namely Kadazans, Dusuns, Ibans, Bidayus and other native groups.Meanwhile, the foreigners, mostly expatriates working and residing in Labuan, were Germans, English, Australians, South Africans and Nigerians.The respondents reported using the internet extensively with 69.6 percent using the internet daily while 25.2 percent using the internet weekly and 5.2 percent using the Internet monthly.Almost 60 percent of the respondents had used Internet for more than 10 years.However, 38.3 percent of the respondents never purchase goods or services on the e-commerce, 41.7 percent had purchase a couple of times in a year and 19.1 percent had purchase monthly on the e-commerce.33 percent of the respondents own 1 credit card, 25 percent own 2 cards, and approximately 41 percents hold 3 or more cards.

Reliability test
Reliability tests were conducted using Cronbach's alpha to measure the stability and internal consistency of the obtained data in this study.As the value is greater than 0.7, the data is deemed reliable and consistent (Nunnally, 1979).Refer to Table 1, the value obtained for all the variables is 0.901.The reliability for trust in the internet structures questions is high, with value of 0.877.The statistic for reliability test for social influences is 0.771 and willingness to purchase online is 0.939.Therefore, it can be said that the overall Cronbach's Alpha values obtained for all variables from the survey data are relatively high, indicating a relatively high probability of stability and internal consistency in the survey questions and the data obtained.

Mean and Standard Deviation of Variables
The mean value of total willingness to purchase online is 2.7983, while the standard deviation is 1.0709.On the scale of 1 to 5, the mean value of willingness to purchase online is slightly below the neutral point.The result indicates that most of the samples are willing to purchase goods and services through the Internet.All facets of trust in the Internet structures mean value is 2.7217 and the standard deviation is 0.8595.On a scale of 1 to 5, the mean value is slightly below the neutral point.This shows that most of the samples trust in the Internet structures.Similar to other variables, all the facets of social influences mean value is slightly below the neutral point.The mean value is 2.4812, while the standard deviation is 0.92422.The result can be concluded as the samples are influences by family members, friends and media in deciding whether to purchase goods or services via Internet.

Regression Analysis
The simple and multiple regression analysis were conducted to determine the relationship between dependent variable of willingness to purchase online and the independent variables of trust in the Internet and social influences.A simple linear regression analysis was used for each independent variable that explained its influence on the dependent separately.However, the multiple linear regression analysis performed is used to combine independent variables which explained its influence on the dependent variable collectively.Meanwhile, the hierarchical regression analysis was used to measure the moderating variable in the independent variable which explained its influence on dependent variable separately with rules included.4.4.1 Hypothesis 1:Trust in the Internet is positively related to willingness to purchase online Trust in the Internet as independent variable was entered into a linear regression model equation with willingness to purchase online as the dependent variable.The result from Table 3 shows that there is significant relationship between trust in the Internet and willingness to purchase online at 1% level.The Durbin-Watson value of 1.904 indicates that there is no autocorrelation problem.The Beta test of slope (ß = 0.876), indicates that there is a highly positive related between trust in the Internet and willingness to purchase online.4.4.2Hypothesis 2: Susceptibility to social influence is positively related to willingness to purchase online Social influence was entered into a linear regression model equation as independent variable, while willingness to purchase online as the dependent variable.From Table 4, the result shows that there is significant relationship between trust in the Internet and willingness to purchase online at 1% level.The Durbin-Watson value of 1.904 indicates that there is no autocorrelation problem.The Beta test of slope (ß = 0.301), indicates that there is a positive related between social influence and willingness to purchase online.4.4.3Hypothesis 3:Susceptibility to social influence is positively related to trust in the Internet Social influence as independent variable was entered into a linear regression model equation, with trust in the Internet trust as the dependent variable.As shown in Table 5, it shows that there is significant relationship between trust in the Internet and trust in the internet at 1% level.The Durbin-Watson value of 1.935 indicates that there is no autocorrelation problem.The Beta test of slope (ß = 0.416), indicates that there is a positive relationship between social influence and trust in the internet.4.4.4Hypothesis 4:Trust in the Internet is positively linked to susceptibility social influence Trust in the Internet as independent variable was entered into a linear regression model equation, with social influence as the dependent variable.From Table 6, it shows that there is significant relationship between trust in the Internet and social influence at 1% level.The Durbin-Watson value of 1.935 indicates that there is no autocorrelation problem.The Beta test of slope (ß = 0.387), indicates that there is a positive relationship between trust in the Internet and social influence.4.4.5 Hypothesis 5:The interaction of the trust in the Internet and susceptibility to social influence is positively related to willingness to purchase online The independent variable of trust in the Internet was entered into hierarchical linear regression with willingness to purchase online as dependent variable and social influence as the moderating variable.Refer the Table 7, the interaction between trust in the Internet and willingness to purchase online moderated by susceptibility to social influence is found to be not significance at 1% level.The result indicates that there is no moderating effect from social influence.

Conclusion
This study was conducted to determine the relationship between trust in the Internet structure and susceptibility to social influence with regards to consumers' willingness to make online purchases in Labuan.The findings support the view that trust in the Internet structure and susceptibility to social influence are significantly related to consumer willingness to purchase online, which are in line with the findings of Lee and Turban (2001) and George (2002George ( , 2004)).
The findings between susceptibility to social influence and willingness to purchase online show a positive relationship.This is consistent with the study conducted by Limayem et. al. (2000) where it was discovered that social influence of family and media was greatly associated with willingness to provide personal information online.Hwang (2005) supported that all three dimensions of social influence (friends, family, media) are significantly related to online trust.In addition, Bhattacherjee (2000) noted that external influence in the form of news reports, popular press and mass media will affect the intention to accept e-commerce.The susceptibility to social influence and trust in the Internet structure were found to be positively related.This result is consistent with the research conducted by Bearden et al. (1989) whereby susceptibility to interpersonal influence was defined as "the tendency to learn about products and services by observing others and seeking information from others".
Hierarchical linear regression analysis had provided insignificant influence between trust in the Internet structure and willingness to purchase online with social influence as the moderator.The result is contradictory to Sinclaire's (2007) study which found that social influence of media as positive and negative messages has a moderating relationship between trust in the Internet and willingness to provide personal information online.However, Sinclaire only measured media as moderator, but this study measured three components which are family members, friends and media as the social influence moderating factor.Thus, this may be the reason why social influence as moderator had no influence as a moderator on trust in the Internet and willingness to purchase online.
The results provided no evidence that trust in the Internet structure of social presence increase users' willingness to purchase online although a previous study discovered that social presence is effective in increasing trust in e-commerce (Gefen & Straub, 2004).The results of the study reported here may be compared to the fear or persuasion relationship proposed by Janis (1967) whereby moderate rather than maximum levels of fear are optimal in effecting persuasion.This study extends current e-commerce studies by investigating social influence as a moderating factor.As such, it investigates the differences in the strength of motivation to purchase online between positive and negative messages of media.Although this study finds social influence is not significant as a moderator between trust in the Internet and willingness to purchase online, it provides meaningful insights for researchers and practitioners.Nevertheless, the e-commerce environment is an ever changing environment.Even though this study shows a consistent result from previous studies conducted, the relationship between variables may change from time to time.This may be due to external factors such as the threats of hackers, resulting in fear in users and dampening their intentions and willingness to purchase online.
For future research, the framework model can be expanded with several additional variables that will influence the willingness to purchase online.Examples include the attractiveness and creativity of the e-commerce website to capture user purchases online, as well as the interactivity of the e-commerce website.