lable at ScienceDirect

Tourism Management 46 (2015) 64e79

Contents lists avai

Tourism Management

journal homepage: www.elsevier .com/locate/ tourman

An integrative model of consumers' intentions to purchase travelonline

Suzanne Amaro a, *, Paulo Duarte b, 1

a Management Department, Polytechnic Institute of Viseu, Campus Politecnico de Repeses, 3504-510 Viseu, Portugalb Business and Economics Department, University of Beira Interior, Estrada do Sineiro - Edifício Ernesto Cruz, 6200-209 Covilh~a, Portugal

h i g h l i g h t s

� An integrative model is proposed to examine determinants of intentions to purchase travel online.� The model integrates constructs of well-established theories of consumer behaviour (TPB, TAM and IDT).� PLS-SEM was used for assessing the measurement model and the structural model.� Findings reveal which variables significantly affect intentions to purchase travel online.� Attitude and compatibility are the most relevant determinants of intentions to purchase travel online.

a r t i c l e i n f o

Article history:Received 3 February 2014Accepted 13 June 2014Available online

Keywords:Innovations Diffusion TheoryIntentions to PurchaseOnline Travel ShoppingSocial MediaTechnology Acceptance ModelTheory of Reasoned ActionTheory of Planned Behaviour

* Corresponding author. Tel.: þ351 232 480 500; faE-mail addresses: samaro@estv.ipv.pt (S. Amaro), p

1 Tel.: þ351 275 319 648; fax: þ351 275 319 600.

http://dx.doi.org/10.1016/j.tourman.2014.06.0060261-5177/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Grounded in the Theory of Reasoned Action, the Theory of Planned Behaviour, the Technology Accep-tance Model and on the Innovation Diffusions Theory, this study proposes and empirically tests an in-tegrated model to explore which factors affect intentions to purchase travel online. Partial Least SquaresStructural Equation Modelling was conducted to assess the hypotheses. The empirical results, obtained ina sample of 1732 Internet users, indicate that intentions to purchase travel online are mostly determinedby attitude, compatibility and perceived risk. The theoretical contributions of this study and the practicalimplications are discussed and future research directions are detailed.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Information Communication Technologies (ICTs) have beenchanging the tourism industry ever since the 1980s (Buhalis& Law,2008), playing a central role in its growth and development(Gretzel & Fesenmaier, 2009). In fact, the establishment of theComputer Reservations Systems and Global Distribution Systems(Sabre, Amadeus, Galileo, andWorldspan) in the 1980s transformedthe tourism industry dramatically. During the late 1980s and early1990s, these systems were important elements for distributingtourism products, with the advantage of providing informationabout customers (Gretzel & Fesenmaier, 2009). However, it was the

x: þ351 232 424 651.duarte@ubi.pt (P. Duarte).

development of the Internet in the 1990s that brought the greattransformation and unprecedented opportunities to the tourismindustry, changing this industry and travellers' behaviour in severalways (Buhalis, 1998; Gretzel & Fesenmaier, 2009).

One of the most significant transformations was that theInternet represented a new and potentially powerful communica-tion and distribution channel for travel suppliers (Law, Leung, &Wong, 2004; Morrison, Jing, O'Leary, & Cai, 2001), fulfilling thegap between consumers and suppliers (Buhalis, 1998). For decades,airlines, cruise lines, the lodging sector and the rental car industryhad been heavily dependent on travel intermediates (e.g. travelagents) to disseminate information and sell their products andservices. With a new distribution channel, these travel suppliersfound away to bypass intermediaries and reach customers directly,while savingmoney (Zhou, 2004). For consumers, the emergence ofthe electronic market brought lower prices, discounts and timesavings (Heung, 2003). In the second half of the 1990s, the

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 65

emergence of online travel agencies, such as Expedia (http://www.expedia.com) and Priceline (http://www.priceline.com) revolu-tionised the way travel was purchased.

The Internet is now an important distribution channel for travel(Lee & Morrison, 2010), generating a worldwide revenue over 340billion United States dollars (USD) in 2011. Worldwide online travelsales have grown 10% each year between 2010 and 2012 and pre-dictions until 2016 show that worldwide online travel sales willcontinue to grow at 8% yearly (eMarketer, 2012).

Given the importance of online travel shopping, the growingbody of literature is not surprising. Yet, in a review of articlespublished in 57 tourism and hospitality research journals from2005 to 2007, Law, Leung, and Buhalis (2009) noted that thenumber of studies related to consumers was relatively small ascompared to the other two categories considered in their study,namely technological development and suppliers. Similarly, afteranalysing information technology in the hospitality industryresearch published between January 2003 and July 2004 in 12hospitality and tourism journals, O'Connor and Murphy (2004)found that “consumer research is largely absent but desperatelyneeded” (p.481). Therefore, the authors suggested themes forfurther research, such as whatmotivates consumers to use a certaindistribution channel and also what motivates them to buy travelonline. On the other hand, as Amaro and Duarte (2013) point out,research addressing online travel shopping presents contradictoryresults and is typically fragmented. The findings of these studiessupport the significance of the current study.

Therefore, this research adopts a distinctive approach to analysethe determinants of intentions to purchase travel online, by pro-posing and empirically testing an integrated model, with contri-butions from well-grounded theories, namely the Theory ofReasoned Action (TRA) (Fishbein & Ajzen, 1975), the Theory ofPlanned Behaviour (TPB) (Ajzen, 1991), the Technology AcceptanceModel (TAM) (Davis, 1985, 1989) and the Innovations DiffusionTheory (IDT) (Rogers, 1995), contributing to the current literaturesince, to the best of knowledge, this has not been done in any otherstudy.

These theories have received substantial empirical support inexplaining users' acceptance in several domains, notably informa-tion systems, and specifically online shopping (e.g. George, 2004;Limayem, Khalifa, & Frini, 2000; Pavlou & Fygenson, 2006; Shim,Eastlick, Lotz, & Warrington, 2001; Yu & Wu, 2007). The pro-posed model seeks to take advantage of the validity, parsimony andreliability that these theories provide as determinants of behaviour,adding other constructs in order to improve explanatory and pre-dictive power. Indeed, besides the constructs derived from thesetheories, trust and perceived risk were added to the model, sinceresearchers have pointed out that their roles in the context of on-line travelling shopping are still unclear (Kim, Chung, & Lee, 2011;Lin, Jones, & Westwood, 2009).

With this integrative approach, it will be possible to determinenot only which variables significantly affect intentions to purchasetravel online, but also the ones that have the strongest impacts,enhancing our understanding of online travel shopping. Addition-ally, the model considers themultidimensionality of two constructs(perceived relative advantages and perceived behavioural control),for a more comprehensive understanding of the factors that influ-ence the purchase of travel online.

The findings will help online travel providers to better un-derstand travellers' online behaviour. Knowing the drivingforces that determine travellers' intentions to purchase travelonline is paramount for the successful implementation of onlinemarketing strategies (Lee, Qu, & Kim, 2007) and to convertpotential customers to actual ones and retain them (Limayemet al., 2000).

2. Literature review

2.1. Intentions to purchase travel online

Intentions to purchase travel online is the main dependentvariable of the model, derived from the TRA, that posits thatbehavioural intentions, rather than attitudes, are the main pre-dictors of actual behaviour (Fishbein & Ajzen, 1975). It is assumedthat intentions capture the motivational factors that influencebehaviour and the stronger the intention to engage in behaviour,the more likely should be its performance (Ajzen, 1991).

Behavioural intentions have been well-established as a strongpredictor of actual usage of information technologies (e.g. Davis,Bagozzi, & Warshaw, 1989; Venkatesh, Morris, Davis, & Davis,2003) and of online shopping (e.g. Ajzen, 2011; Limayem et al.,2000; Lin, 2007; Pavlou & Fygenson, 2006). Furthermore, involuntary settings, as in the case of online travel shopping, inten-tion to behave has been postulated as the best predictor ofbehaviour (Moital, Vaughan, & Edwards, 2009).

2.2. Attitude towards online travel shopping

According to the TRA, intentions are the result of attitudes to-wards the outcomes of behaviour (Fishbein & Ajzen, 1975). InFishbein and Ajzen's (1975) seminal work, the researchers discussthe confusion and ambiguity surrounding the attitude concept,namely because of the wide range of existing definitions and mea-sures. However, there seems to be a widespread agreement thataffect, defined as a person's feelings towards an object, person, issueor event, is the most essential part of the attitude concept (Fishbein& Ajzen, 1975). According to Fishbein and Ajzen (1975), attitude is a“learned predisposition to respond in a consistently favourable orunfavourable manner with respect to a given object” (p.15).

It is expected that positive attitudes will lead to higher in-tentions to perform the behaviour. Following Fishbein and Ajzen's(1975) definition, in this study attitude is defined as the strengthof a person's feeling of favourableness or unfavourableness towardsthe purchase of travel online.

In the travel context, several studies have evidenced that atti-tude towards online shopping positively influences intentions topurchase travel online (Bign�e, Sanz, Ruiz, & Ald�as, 2010; Lee et al.,2007; Morosan & Jeong, 2008). Additionally, among the variablesBign�e et al. (2010) included in their study, attitude had the stron-gest effect on intentions to purchase travel online.

Therefore, as intention is determined by the person's positive ornegative attitudes towards the decision, it is expected that:

H1: Individuals' attitude towards online travel shopping posi-tively influences intentions to purchase travel online.

2.3. Communicability

According to the TRA a person's behavioural intention is not onlya function of attitude towards the behaviour, but also of subjectivenorm, described as the person's perceptions of the social pressuresto perform the behaviour (Ajzen,1985; Fishbein& Ajzen,1975). As ageneral rule, if individuals believe that their referents think theyshould perform the behaviour, than they will perceive socialpressure to do so (Ajzen, 1985).

While subjective norm may explain intentions to performcertain behaviours, several authors have also evidenced that sub-jective norm was neither significant in predicting intentions topurchase online (e.g. Lin, 2007; Pavlou & Fygenson, 2006; Wang,Chen, Chang, & Yang, 2007) nor actual online purchases (e.g.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7966

George, 2004). More recently, a study in the context of online travelshopping evidenced that social influence did not affect onlinepurchase intentions (San Martín & Herrero, 2012). Venkatesh andDavis (2000) argue that subjective norm only has an effect on in-tentions in mandatory usage contexts, and not when the usage isvoluntary. However, it is undeniable that the most pervasive factorsthat influence an individual's behaviour is the influence of others(Burnkrant & Cousineau, 1975). Therefore, due to the importance ofconsidering the influence of others and since prior work has foundthat subjective norm has not performed well in explaining in-tentions to purchase online, this study suggests employingcommunicability, a different form of social influence. Communica-bility is related to the influence of family and friends, in the sensethat people are more likely to book online and to frequently booktravel online if they know that other people are doing likewise(Morrison et al., 2001). Yet, there is no general consensus on thismatter. Indeed, Li and Buhalis (2006) asserted that communica-bility was not important in explaining the adoption of online travelshopping. To further explore this inconsistency the following hy-pothesis is formulated:

H2: Communicability positively influences intentions to pur-chase travel online.

Kim, Qu, and Kim (2009) also found that the recommendation offamily and friends was important to reduce the risk perceived withonline travel purchases. Thus, knowing that families and friendspurchase online can relieve customer's anxiety in purchasing onlineand reduces perceived risk (Corbitt, Thanasankit, & Yi, 2003), it isproposed that:

H3: Communicability negatively influences perceived risk ofintentions to purchase travel online.

2.4. Perceived complexity

The Innovations Diffusion Theory (IDT), originally developed byEverett Rogers in 1962, is one of the most frequently used theoriesto explain technological innovation (Hung, Yang, Yang, & Chuang,2011). According to Rogers (1995), five innovation characteristicse Perceived Complexity, Perceived Compatibility, Perceived Rela-tive Advantages, Trialability and Observability2 e will determine ifadoption or diffusion will occur.

Perceived complexity reflects the extent inwhich it is difficult tounderstand and, consequently, to use (Rogers, 1995). As one canexpect, innovations that are simpler to understand will be adoptedmore rapidly than innovations that require new skills (Rogers,1995). For the purpose of this study, perceived complexity is thedegree towhich purchasing travel online is perceived to be difficult.

The measurement scales and definition of complexity isconsiderably similar to TAM's perceived ease of use (Venkateshet al., 2003). Several researchers have highlighted these similar-ities (Davis, 1989; Moore & Benbasat, 1991; Wu & Wang, 2005).Common sense and theory suggest that innovative technologiesthat are perceived to be easier to use and less complex have ahigher possibility of acceptance and use by potential users (Daviset al., 1989; Shih & Fang, 2004). Therefore, the TAM posits thatease of use is a determinant of attitude (Davis, 1989). In the travelcontext, studies grounded on the TAM have found that perceivedease of use affected consumer's attitude towards online travel

2 Only Perceived Complexity, Perceived Compatibility and Perceived RelativeAdvantages were included in this study, as only these were found to be consistentlyrelated to adoption (Tornatzky & Klein, 1982).

agencies (Cho & Agrusa, 2006) and on attitude towards using hotelreservation websites (Morosan & Jeong, 2008).

Based on these findings and on the TAM, it is expected that theperceived complexity of online travel shopping will be a determi-nant of attitude towards online shopping. Thus:

H4: Individuals' perceived complexity of online travel shoppingwill be negatively related to attitude towards online travelshopping.

2.5. Perceived compatibility

Perceived compatibility is a construct borrowed from the IDTdefined as “the degree to which an innovation is perceived as beingconsistent with existing values, past experiences, and needs ofpotential adopters” (Rogers, 1995, p. 15). Based on Vijayasarathy's(2004) definition of compatibility applied to online shopping, forthe purpose of the current study, compatibility is the extent towhich consumers believe that purchasing travel online fits/matches their lifestyle, needs, and shopping preference.

Research has supported the positive and significant relationshipbetween compatibility and attitude towards online shopping (e.g.Chen, Gillenson, & Sherrell, 2002; Vijayasarathy, 2004). Bellman,Lohse, and Johnson (1999) reported that individuals who spent aconsiderable amount of time using the Internet and other relatedtechnologies such as e-mail in their job or personal life would bemore likely to shop online. Christou and Kassianidis (2003) and Liand Buhalis (2006) also found that perceived compatibility waspositively associated with intentions to purchase travel online.Thus, it is hypothesised that:

H5: Individuals' perceived compatibility with online travelshopping will be positively related to attitude towards onlinetravel shopping.

H6: Individuals' perceived compatibility with online travelshopping will be positively related to intentions to purchasetravel online.

2.6. Perceived behavioural control

Perceived behavioural control was added to the TRA to over-come the original model's limitations in dealing with behavioursover which people have incomplete volitional control (Ajzen,1991),creating the Theory of Planned Behaviour. Indeed the TRA canpredict behaviours from intentions with a high degree of accuracy,provided that the behaviours are under volitional control (Ajzen &Fishbein, 1980), otherwise behaviours may not be performed.When people believe that they have the resources and opportu-nities and that the obstacles theymay encounter can be overcomed,then they shall have the confidence to perform the behaviour, andtherefore exhibit a high degree of perceived behavioural control(Ajzen, 2002).

Some researchers have argued that perceived behavioural con-trol is not well understood (e.g. Pavlou & Fygenson, 2006;Trafimow, Sheeran, Conner, & Finlay, 2002), with studies employ-ing self-efficacy instead of perceived behavioural control whenconducting research grounded on the TPB (e.g. Krueger, Reilly, &Carsrud, 2000; Vijayasarathy, 2004). Although self-efficacy andperceived behavioural control are related concepts, they cannot beused interchangeably and should be distinguished (Terry, 1993).Self-efficacy is related with cognitive perceptions of control basedon internal factors, while perceived behavioural control reflectsboth internal and external factors (Armitage & Conner, 2001).

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 67

More than 10 years after perceived behavioural control wasadded to the TRA Ajzen (2002), aware of the problems with thenature andmeasurement of this construct, explicitly recommendedthe decomposition of perceived behavioural control in two com-ponents: self-efficacy and controllability. While self-efficacy refersto ease or difficulty of performing a behaviour or confidence inone's ability to perform it, controllability refers to control over thebehaviour, or the beliefs about the extent to which performing thebehaviour is up to the actor (Ajzen, 2002).

The current study follows these recommendations and de-composes perceived behavioural control in two components: self-efficacy and controllability. This structure maintains the parsimo-nious unitary view of perceived behavioural control and provides amore detailed prediction of external control beliefs by allowing adistinct prediction of self-efficacy and controllability, which willlead to a better prediction of perceived behavioural control,intention and behaviour (Pavlou & Fygenson, 2006).

For the purpose of this study, the following definitions will beconsidered:

� Online travel shopping self-efficacy: Following Pavlou andFygenson (2006) and Vijayasarathy (2004), online travel shop-ping self-efficacy is defined as consumers' self-assessment oftheir own capabilities to purchase travel online.

� Controllability: Grounded on Ajzen's (2002) and Pavlou andFygenson's (2006) definitions, controllability is defined as in-dividual judgements about the availability of resources andopportunities to purchase travel online.

The relationship between perceived behavioural control andintentions to purchase travel online has clearly been underlookedat. Bign�e et al.'s (2010) study was the only one that used theperceived behavioural control construct and found that it did notdirectly influence users' intention to purchase airline tickets online.In a different study, that considered facilitating conditions, aconcept similar to perceived behavioural control, San Martín andHerrero (2012) found that they failed to predict intentions tobook a room online.

Grounded on the TPB, that conceptualises that perceivedbehavioural control is held to contribute to intentions (Ajzen,1991),this study proposes the following hypothesis:

H7: Individuals' perceived behavioural control over purchasingtravel online positively influences intentions to purchase travelonline.

Hernandez, Jimenez, and Martín (2009) demonstrated that in-dividuals who felt they had the capability of purchasing onlinewould perceive online shopping as easier to use. Hence, it ishypothesised that:

H8: Individuals' perceived behavioural control over purchasingtravel online negatively influences perceived complexity.

2.7. Perceived relative advantage

One of the IDT core constructs is relative advantages, a conceptsimilar to TAM's perceived usefulness. In fact, many researchersconsider that they are equivalent (Lee & Kozar, 2008;Riemenschneider, Hardgrave, & Davis, 2002; Wu & Wang, 2005).However, there exists an important distinction between the twoconcepts since relative advantages explicitly contain a comparisonbetween the innovation and its precursor, while perceived useful-ness does not (Karahanna, Ahuja, Srite, & Galvin, 2002; Shin, 2010).

Since the current study is interested in understanding users' per-ceptions of the advantages of online travel shopping over tradi-tional channels, relative advantage is considered to be moreadequate than perceived usefulness, since it is a broader concept.

For the purposes of the current study, relative advantage isdefined as the degree to which online travel shopping providesbenefits to consumers or is better than its alternatives, such aspurchasing at high street travel agencies or directly contactingtravel suppliers by telephone or fax.

From the literature, several major issues emerge as advantagesof online travel shopping and have typically included convenience(Heung, 2003; Jensen, 2009; Kim & Kim, 2004; Kim, Ma, & Kim,2006; Kolsaker, Lee-Kelley, & Choy, 2004; Mayr & Zins, 2009),financial advantages, such as lower prices (Kim & Kim, 2004; Kim,Kim, & Han, 2007; Kim et al., 2006), time saving (Christou &Kassianidis, 2003; Heung, 2003; Wong & Law, 2005), enjoyment(Cho & Agrusa, 2006; Powley, Cobanoglu, & Cummings, 2004) andproduct variety (Jensen, 2009).

The current study considers that these are the pertinent di-mensions of relative advantage, because they represent ways inwhich online travel shopping can offer advantages over traditionalchannels. Hence, relative advantages of online shopping are con-ceptualised as a multidimensional construct that captures thesebenefits of online shopping. This overall abstraction is believed tobe theoretically meaningful and parsimonious to use as a repre-sentation of the dimensions (Law, Chi-Sum, & Mobley, 1998).

The perceived relative advantages of online travel shoppinghave been found to affect intentions to purchase online (Christou&Kassianidis, 2003; Kim & Kim, 2004; Kim et al., 2006; Moital,Vaughan, Edwards, & Peres, 2009; Wong & Law, 2005) and alsoinfluences the adoption of online travel shopping (Heung, 2003;Jensen, 2009; Kamarulzaman, 2007; Morrison et al., 2001).

Based on these arguments, the following hypothesis is posited:

H9: Perceived relative advantages of online travel shopping willbe positively related to intentions to purchase travel online.

Additionally, grounded on the TAM that suggests the perceivedusefulness (that is considered to be integrated in the relative ad-vantages construct) affects attitude, it is hypothesised that:

H10: Perceived relative advantage of online travel shopping willbe positively related to attitudes towards online travel shopping.

To the best of our knowledge, a relationship that has never beenexplored in online shopping is the one between perceived relativeadvantages and trust. In the context of Internet banking, SuhandHan(2002) found that customers' perceived usefulness had a positiveimpact on trust in Internet banking. Based on thisfinding, it is arguedthat individualswhoperceive the relative advantages of online travelshopping are more likely to trust online shopping and therefore:

H11: Perceived relative advantage of online travel shopping willincrease trust in online travel shopping.

2.8. Perceived risk

The present study is concerned in examining the perceived riskwith the Internet as the purchase method for travel and not withthe travel service itself. Accordingly, perceived risk is defined as thepotential loss perceived by a consumer in considering the purchaseof travel online when compared to the purchase of travel offline.

Surprisingly, very little research has looked at perceived riskassociated with online travel shopping (Lin et al., 2009). Studies

3 The second-order constructs were operationalised using the repeated indicatorapproach as it produces more precise parameter estimates and a more reliablehigher-order construct score for reflective-formative hierarchical constructs(Becker, Klein, & Wetzels, 2012).

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7968

have evidenced that perceived risk is negatively related to thepurchase of airline tickets online (Kim, Kim, & Leong, 2005; Kimet al., 2009; Kolsaker et al., 2004). Considering not just airlinetickets but all types of travel services, Jensen (2012) found thatperceived risk was negatively related to consumers' intention topurchase travel online.

Since research has attested that perceived risk has a negativeeffect on intentions topurchase travel online (Jensen, 2012; Kolsakeret al., 2004) and on attitude towards online travel shopping (Bign�eet al., 2010) the following research hypotheses are proposed:

H12: The perceived risk of online travel shopping has a negativeinfluence on attitude towards online travel shopping.

H13: The perceived risk of online travel shopping has a negativeinfluence on intentions to purchase travel online.

2.9. Trust

Trust in online travel shopping is defined as “an attitude ofconfident expectation in an online situation of risk that one's vul-nerabilities will not be exploited” (Corritore, Kracher, &Wiedenbeck, 2003, p. 740). Research has shown that people aremore prone to purchase online if they perceive a higher trust inonline shopping (e.g. Corbitt et al., 2003). Indeed, trust plays acentral role on online purchases, because consumers will hesistateto purchase if they feel uncertainty and risk (McKnight, Choudhury,& Kacmar, 2002). In a more extreme view, Wang and Emurian(2005) posit that the future of online shopping depends on trust.

Despite the importance of trust, Kim et al. (2011) state that thereis a lack of research regarding perceived trust in online shopping fortourism products and services. Therefore, it is relevant to add trustsince the few studies that have considered trust in online travelshopping have also produced mixed results. For example, whileWen (2010) claimed that consumers' trust in online shopping had apositive effect on intentions to purchase travel online,Kamarulzaman (2007) did not find a direct effect on the adoption ofonline travel shopping. To clarify these mixed results, the followinghypothesis was set forth:

H14: Trust in online travel shopping has a positive influence onintentions to purchase travel online.

Other research has found that trust influences attitude towardsonline shopping (Bign�e et al., 2010) and negatively influencesperceived risk (Kamarulzaman, 2007). Thus, it is hypothesised that:

H15: Trust in online travel shopping has a positive influence onattitude towards online shopping.

H16: Trust in online travel shopping has a negative influence onperceived risk of online travel shopping.

Fig. 1 graphically summarises the research hypotheses.

3. Methodology

3.1. Sampling and data collection

Theoretically, the population comprises all Internet users aged18 or more as they have already purchased travel online or exhibit agreater propensity to shop online. However, since there does notexist a list of Internet users it is impossible to select our samplingelements from the population directly. Consequently, a non-probabilistic sampling procedure e convenience sampling e wasused to collect data.

To collect the data the questionnaire was distributed online,using the Survey Gizmo questionnaire tool (www.surveygizmo.com). Since the study focuses on online travel shopping, it wasnot necessary to address the concerns of individuals that do nothave access to the Internet.

In late July of 2012, e-mail invitations were sent to colleagues,students, personal contacts, professional list-serve lists, and othere-mail contacts collected. The e-mail invitations provided re-spondents with information on the purpose of the study, theapproximate time to fill out the questionnaire and information onthe prize draw for thosewho completed the survey. Moreover, linksto the survey were placed on Facebook and LinkedIn.

These contacts and list-serve lists were composed majority ofPortuguese Internet users, but also Internet users from all over theworld. Therefore, the questionnaire was available in Portugueseand in English. Each language had its own URL link, although it waspossible to switch languages on the first page of the questionnaire.

The questionnaire was available online between July 17th andSeptember 12th of 2012. During this period, a total of 1759 re-sponses was obtained. Since the total number of responses waslarge, the complete case approach was used (Hair, Black, Babin, &Anderson, 2010) and responses with missing values were elimi-nated. Thus, a total of 1732 responses was considered valid forfurther analyses. It should be stressed that the most common ruleto determine sample size for PLS estimation consists in deter-mining the sample according to the most complex multipleregression in the model, which consists in either the number ofindicators on the most complex formative construct or the largestnumber of antecedents leading to a construct in the inner model(Barclay, Higgins, & Thompson, 1995). Once determined which isgreater, the sample size required is 10 cases per predictor. In theproposedmodel, the most complex regression involves the numberof structural paths directed at the intentions to purchase travelonline construct, which are seven. Thus, according to this rule, theminimum sample size necessary would be 70. With 1732 re-sponses, the PLS analysis appears to have sufficient power.

3.2. Questionnaire and measurements

The questionnaire for the current study was divided into 2 mainsections. The first section contained questions directed to the scaleitems (indicators), selected to measure each construct based onexistingmeasures or adapted from similar scales (see Appendix 1 forthe itemsusedand sources). All itemsweremeasuredusinga5-pointLikert scale,with 5 being “Strongly Agree” and 1 “Strongly Disagree”,except for item INT1, in which 5 was “Very High” and 1 “Very Low”.

It should be noted that all first-order constructs have a reflectivemeasurement, where the indicators are considered to be functionsof the latent construct (Hair, Ringle, & Sarstedt, 2011; Hair et al.,2010). The second-order constructs (perceived relative advan-tages and perceived behavioural control), however, have a forma-tive measurement, since the first-order variables are assumed tocause the second-order variables, i.e., changes in the first-ordervariables would cause changes in the underlying variable (Jarvis,MacKenzie, & Podsakoff, 2003).3

The last section of the questionnaire contained questionsregarding respondents' demographic characteristics, namely age,gender and education level. In this section, respondents were alsoasked to leave their e-mail contact in case they wanted to enter the

Fig. 1. The theoretical model.

Table 1Demographic profile of respondents.

Variable Category N % of Responded

Age 18e29 599 34.630e39 496 28.640e49 404 23.350e59 179 10.3Over 60 54 3.1TOTAL 1732 100

Gender Male 667 38.5Female 1065 61.5TOTAL 1732 100

Education Level 12th grade or less 201 11.6College Degree 565 32.6Master Degree 576 33.3Doctoral Degree 390 22.5TOTAL 1732 100

Continent of Residence Asia 52 3Africa 19 1.1Europe 1531 88.39North America 27 1.56South America 84 4.85Oceania 19 1.1TOTAL 1732 100

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 69

prize draw of an Amazon voucher and a free night at a 5 star hotel,offered in order to increase response rates. To prevent duplicateresponses, the option to control and remove duplicate responses byIP address available in the Surveygizmo platform was used. Theonline questionnaire listed only a few questions per screen, as thereis some evidence that suggests that this procedure minimisesabandonment (Schonlau, Fricker, & Elliott, 2002).

3.3. Data analysis

Partial least squares structural equation modelling (PLS-SEM),using the SmartPLS 2.0 programme (Ringle, Wende, & Will, 2005),was used to validate the measures developed and test the hy-potheses. This approach readily incorporates both reflective andformative measures and has less restrictive assumptions about thedata (Hair, Ringle, Hult, & Sarstedt, 2013; Hair, Sarstedt, Pieper, &Ringle, 2012; Hair et al., 2011). For instance, PLS does not requirea normal distribution since it uses bootstrapping to empiricallyestimate standard error for its parameter estimates (Gefen, Rigdon,& Straub, 2011; Henseler, Ringle, & Sarstedt, 2012). Therefore,normality in the distribution was not checked for. Moreover, con-structs with fewer items can be used. Since five constructs in themodel (Intentions to purchase travel online, Perceived Compati-bility, Enjoyment, Self-efficacy and Controllability) only have twoitems, this characteristic seemed relevant.

4. Results

4.1. Socio-demographic characteristics

A demographic profile of survey participants is summarised inTable 1. The age group with the most significant number of

responses was the age group 18e29, with 34.6% of the total of re-sponses, while only approximately 13% are aged over 50.

In terms of gender, there is a slight skew towards a higherproportion of female participants (61.5%). The sample seems to becomposed by highly educated individuals, with approximately 88%of the respondents holding at least a college degree, against only11.6% who have only completed the 12th grade or less. Regardingthe country of residence, therewas a prominence of responses fromEuropean residents.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7970

4.2. Model assessment

The evaluation of a researchmodel using PLS analysis consists oftwo distinct steps. The first step includes the assessment of themeasurement (outer) model and deals with evaluating the char-acteristics of the constructs and measurement items that representthem. The second step involves the assessment of the structural(inner) model and the evaluation of the relationships between theconstructs as specified by the research model. It should be notedthat PLS path modelling does not provide any global goodness-of-fit criterion, as CB-SEM does, which implies a lack of measures foroverall model fit (Hair, Sarstedt, Ringle, & Mena, 2012; Henseleret al., 2012).

4.2.1. Measurement modelThe first part in evaluating a model is to present the outer model

results to examine the reliability and validity of the measures usedto represent each construct (Chin, 2010). However, procedures usedto assess the validity and reliability of reflective constructs aredifferent than the ones used to assess formative constructs(Diamantopoulos & Winklhofer, 2001). Since the model has bothreflective and formative measurements, the measurement modelassessment will be divided to distinguish these two types ofconstructs.

4.2.1.1. Reflective measurement. Assessment of reflective constructsinvolves determining indicator reliability, internal consistencyreliability, convergent validity and discriminant validity, asdescribed by Hair et al. (2011) and Henseler, Ringle, and Sinkovics(2009).

As shown in Table 2, all measures are robust in terms of theirreliability, since all Cronbach's alpha are higher than 0.7, except forControllability with 0.628, but still above the acceptable thresholdof 0.6 (Hair et al., 2010). Furthermore, the composite reliabilities,that many researchers consider more suitable for PLS-SEM thanCronbach's alpha (e.g. Hair et al., 2011; Henseler et al., 2009), rangefrom 0.84 to 0.97, which exceed the recommended threshold valueof 0.70 (Bagozzi & Yi, 1988). Finally, all indicator loadings are abovethe 0.6 cut-off (Chin, 1998; Henseler et al., 2009).

To evaluate convergent validity, as suggested by Fornell andLarcker (1981), each construct's AVE was calculated (see Table 2).The results support convergent validity, since they all exceed 0.50,ranging from 0.57 to 0.89.

Discriminant validity was assessed following Fornell andLarcker (1981) guidelines, to examine if a construct is morestrongly related to its own measures than with any other constructby examining the overlap in variance by comparing the AVE of eachconstruct with the squared correlations among constructs (Chin,2010). Table 3 shows the correlations between constructs, wherethe diagonal elements are the square roots of the AVEs. Asobserved, the square root of each construct's AVE is larger than itscorrelations with any other construct. Therefore, each constructshares more variance with its own block of indicators than withanother latent variable representing a different block of indicators(Henseler et al., 2009), supporting the adequate discriminant val-idity of the scales.

Discriminant validity was further assessed by extracting thefactor and cross loadings of all indicators to their respective con-structs. Not only should each indicator be strongly related to theconstruct it attempts to reflect, but also should not have a strongerconnection with another construct (Chin, 2010). The results, pre-sented in Appendix 2, indicate that all indicators loaded on theirrespective constructmore highly than on any other, confirming thatthe constructs are distinct.

In conclusion, these results provide support for the overallquality of the reflective constructs' measures.

4.2.1.2. Formative constructs. The formative constructs proposed inthe model e perceived behavioural control and perceived relativeadvantages - are second-order constructs. Tests of measurementquality for a second-order factor model should, by analogy, followthe same process that is used to examine the first-order factors(Chin, 1998). Therefore, the assessments of measurement quality ofsecond-order constructs are conducted in two stages, at the first-order construct level (carried out in the previous section, since allfirst-order constructs are reflective) and at the second-orderconstruct level, in which the first-order constructs act as indicatorsof the second-order construct (Hair et al., 2010).

The weights of the first-order constructs on the second-orderconstructs and their significance were examined (see Table 4), toassess if each first-order construct contributes to form the sec-ond-order construct (Chin, 1998; Hair et al., 2011). For a forma-tive higher-order construct, the weights of the lower-orderconstructs are especially important as they represent actionabledrivers of the higher-order construct (Becker, Klein, & Wetzels,2012).

All first-order constructs weights are significant, which meansthat there is empirical support for the first-order constructs rele-vance for the construction of the formative second-order constructsas theoretically conceived, demonstrating a sufficient level of val-idity (Hair et al., 2011; Urbach & Ahlemann, 2010). Moreover, theweights are higher than 0.10 and their sign is consistent with theunderlying theory (Andreev, Heart, Maoz, & Pliskin, 2009).Regarding perceived relative advantages, it should be noted thatfinancial advantages, time saving and convenience are the mostrelevant advantages.

Another important criterion for assessing the validity of thefirst-order constructs is to examine multicollinearity. Unlike con-structs with a reflective measurement, where multicollinearitybetween construct items is desirable, excessive multicollinearitybetween the formative first-order constructs can destabilise themodel (Diamantopoulos & Winklhofer, 2001) and may cause theweights to be non-significant and, therefore, redundant (Hair et al.,2011). If the first-order constructs are highly correlated, it maysuggest that they are tapping into the same aspect of the construct(Petter, Straub, & Rai, 2007) and therefore, a formative nature forthe second-order construct would be inappropriate. Therefore, toensure that multicollinearity was not present, the variance inflationfactor (VIF) was determined, with values varying from 1.504 to amaximum of 3.083, which is far below the common cut-offthreshold of 5 (Hair et al., 2011).

At the second-order construct level it is important to assess thenomological validity, i.e., if the formative construct carries theintended meaning. This may be manifested in the magnitude andsignificance of the relationships between the second-orderformative construct and other constructs in the research model,which are expected to be strong and significant based on previousresearch (Henseler et al., 2009).

Table 5 shows the relationships between the second-orderconstructs and other constructs in the model, according to thehypotheses proposed in the model.

The results indicate significant relationships between perceivedbehavioural control and perceived relative advantages with otherconstructs in the model, consistent with underlying theory, indi-cating nomological validity. Although perceived relative advantagedoes not exhibit a significant effect on intentions to purchase travelonline, as hypothesised, this is due to the mediating effect of atti-tude, as will further be discussed.

Table 2Measurement statistics of construct scales based on reflective indicators.

Construct/Indicators Mean StandardDeviation

IndicatorLoadings

t-valuea Compositereliability

Cronbach’s a Average VarianceExtracted

Intentions to Purchase Travel Online 3.95 1.03 0.94 0.87 0.89INT1 3.89 1.17 0.94 278.60INT2 4.00 1.02 0.94 199.97

Attitude 3.96 0.78 0.94 0.91 0.75ATT1 4.10 0.82 0.90 143.17ATT2 4.01 0.90 0.94 243.34ATT3 3.89 0.92 0.91 142.33ATT4 3.89 0.90 0.90 102.24ATT5 3.85 0.99 0.66 27.90

Communicability 4.15 0.65 0.88 0.79 0.71CMM1 4.29 0.72 0.85 60.06CMM2 4.21 0.76 0.90 131.00CMM3 3.88 0.86 0.77 48.12

Perceived Complexity 2.09 0.64 0.84 0.75 0.57CXY1 2.37 1.02 0.69 30.56CXY2 2.08 0.85 0.79 51.66CXY3 1.88 0.72 0.83 80.35CXY4 2.15 0.87 0.70 37.98

Compatibility 3.66 0.95 0.93 0.84 0.86CMP1 3.79 0.99 0.94 321.47CMP2 3.50 1.07 0.92 155.73

Self-Efficacy 4.15 0.78 0.93 0.85 0.87SEF1 4.11 0.86 0.93 174.95SEF2 4.20 0.82 0.94 199.64

Controllability 4.38 0.65 0.84 0.63 0.73CTR1 4.51 0.64 0.87 140.99CTR2 4.17 0.97 0.83 72.22

Convenience 4.08 0.62 0.85 0.73 0.65CNV1 4.24 0.73 0.79 61.08CNV2 3.88 0.82 0.79 64.80CNV3 4.10 0.75 0.84 76.99

Financial Advantages 3.79 0.72 0.91 0.85 0.77FAD1 3.91 0.84 0.89 160.01FAD2 3.72 0.83 0.90 142.68FAD3 3.74 0.80 0.84 76.69

Time Saving 3.98 0.71 0.91 0.86 0.78TSV1 4.03 0.77 0.91 155.53TSV2 4.07 0.76 0.92 163.84TSV3 3.80 0.88 0.82 68.63

Enjoyment 3.06 0.83 0.93 0.84 0.86EJY1 3.06 0.88 0.92 151.34EJY2 3.06 0.91 0.94 254.61

Product Variety 3.48 0.74 0.87 0.78 0.69PVR1 3.40 0.87 0.81 62.69PVR2 3.51 0.87 0.86 95.27PVR3 3.55 0.93 0.83 99.46

Perceived Risk 2.70 0.83 0.90 0.96 0.64RSK1 2.50 0.94 0.83 89.39RSK2 2.76 1.13 0.80 74.45RSK3 2.64 1.08 0.80 61.57RSK4 2.82 1.02 0.75 50.87RSK5 2.80 1.03 0.81 71.55

Trust 3.55 0.66 0.87 0.81 0.57TRT1 3.29 0.92 0.66 33.91TRT2 3.53 0.79 0.65 33.03TRT3 3.61 0.79 0.82 89.01TRT4 3.49 0.95 0.83 75.92TRT5 3.71 0.89 0.81 66.40

a t-values were obtained with the bootstrapping procedure (5000 samples) and are significant at the 0.001 level.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 71

4.2.2. Inner model assessmentSince the outer model evaluation provided evidence of reli-

ability and validity, the inner model estimates were examined toassess the hypothesised relationships among the constructs in theconceptual model (Hair et al., 2013). The inner model proposed inthis studywas evaluatedwith several measures, following Henseleret al.'s (2009, 2012) and Hair et al.'s (2013) recommendations.

The standardised path coefficients and significance levels pro-vide evidence of the inner model's quality (Hair, Sarstedt, Ringleet al., 2012) and allow researchers to test their proposed

hypotheses. The path coefficients and significance levels are illus-trated in Fig. 2.

The total effects (indirect effect þ direct effect) of the inde-pendent constructs on the dependent ones were also examined,since they provide practitioners with actionable results regardingcause-effect relationships. Table 6 shows the direct, indirect andtotal effects of the predictors of the main dependent variable of themodel, intentions to purchase travel online.

The first hypothesis predicted that attitude towards onlinetravel shopping would positively influence intentions to purchase

Table 3Discriminant validity of the constructs e correlations between constructs.

Constructs 1 2 3 4 5 6 7 8 9 10 11 12 13 14

1. Attitude 0.872. Communicability 0.49 0.843. Compatibility 0.74 0.33 0.934. Complexity �0.5 �0.34 �0.52 0.765. Controllability 0.43 0.34 0.44 �0.54 0.856. Convenience 0.58 0.39 0.59 �0.55 0.5 0.87. Enjoyment 0.42 0.18 0.44 �0.24 0.14 0.36 0.938. Financial Advantage 0.51 0.29 0.47 �0.37 0.29 0.58 0.38 0.889. Perceived Risk �0.52 �0.27 �0.58 0.56 �0.36 �0.46 �0.28 �0.34 0.810. Product Variety 0.44 0.24 0.45 �0.27 0.2 0.45 0.6 0.53 �0.28 0.8311. Purchase Intentions 0.78 0.39 0.7 �0.46 0.43 0.51 0.33 0.44 �0.53 0.39 0.9412. Self-Efficacy 0.53 0.32 0.57 �0.59 0.58 0.49 0.24 0.31 �0.49 0.24 0.53 0.9313. Time Saving 0.51 0.34 0.46 �0.43 0.37 0.63 0.39 0.52 �0.34 0.46 0.41 0.35 0.8814. Trust 0.59 0.32 0.63 �0.55 0.38 0.52 0.35 0.41 �0.73 0.35 0.53 0.49 0.42 0.76

Bold numbers represent the square roots of the AVEs.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7972

travel online. Consistent with intention based models and withother studies conducted in the online travel context (e.g. Bign�eet al., 2010; Lee et al., 2007; Morosan & Jeong, 2008), attitudewas found to be significantly associatedwith intentions to purchasetravel online (b ¼ 0.55, p < 0.001).

The second and third hypotheses proposed that communica-bility influences intentions to purchase travel online and perceivedrisk. The former hypothesis was based on Morrison et al.'s (2001)findings that people were more likely to book online if they knewthat many other people were doing likewise. However, this hy-pothesis was not supported (b ¼ 0.01, p ¼ 0.33). This result is nottotally unexpected, since the influence of others to perform be-haviours in volunteering settings e such as the purchase of travelonline e has been found to be weak or non-existent (e.g. Daviset al., 1989; Hsu & Chiu, 2004; San Martín & Herrero, 2012; Shih& Fang, 2004). The current results support Li and Buhalis's (2006)findings that communicability is not an important factor inexplaining the purchase of travel online. The latter hypothesispertaining that communicability would reduce people's perceivedrisk regarding the purchase of travel online was also not supported(b¼�0.03, p¼ 0.09). Both hypotheses using communicability wererejected leading us to interesting conclusions. It suggests that theinfluence of friends tends to diminish as the purchase of travelonline gets more widespread and that it is most likely that thisinfluence is important for new phenomenon.

The fourth hypothesis that stated “Individuals' perceivedcomplexity of online travel shopping will be negatively related toattitude towards online travel shopping”, was supported(b ¼ �0.05, p < 0.05). This finding is right in line with the IDT thatposits that innovations that are simpler to understand will beadopted more quickly (Rogers, 1995). Purchasing travel online canbe a complex process, since there are many travel suppliers onlineand different procedures to conclude the purchase. Internet usersthat feel that it is complex will have a less favourable attitude to-wards the purchase of travel online. However, this effect is

Table 4Weights of the first-order constructs on the second-order constructs.

2nd Order Constructs 1st Order Constructs Weight t-Statistic

Perceived Behavioural Control Self-Efficacy 0.64 60.79***Controllability 0.48 61.01***

Perceived Relative Advantages Convenience 0.27 39.38***Financial Advantage 0.30 41.94***Time Saving 0.30 42.32***Enjoyment 0.18 30.27***Product Variety 0.24 37.39***

***Significant at 0.001 level based on 5000 bootstraps.

relatively weak. Complexity appears to be relevant when an indi-vidual first starts a new behaviour, but after a period of time itsinfluence becomes less significant (Vijayasarathy, 2004).

Internet users' perception about the compatibility of onlinetravel shopping appears to be a strong predictor of online travelshopping. Indeed, the fifth and sixth hypothesis predicting a rela-tionship with attitude and intentions to purchase travel online,respectively, were supported (b ¼ 0.47, p < 0.001 and b ¼ 0.20,p < 0.001). This result suggests that people who feel that onlinetravel shopping is compatible with their lifestyle will have a morefavourable attitude towards online travel shopping and can be ex-pected to purchase travel online more readily.

The seventh and eighth hypotheses were concerned with therole of perceived behavioural control, a construct borrowed fromthe TPB. The seventh hypothesis was supported by the data(b ¼ 0.10, p < 0.001), indicating that perceived behavioural controlpositively influences intentions to purchase travel online, echoingthe postulation of the TPB. This result helps to better clarify thecontradictory finding of a previous study regarding the purchase oftravel online (Bign�e et al., 2010) that had found that perceivedbehavioural control did not affect intentions to purchase travelonline.

Data analysis also indicates that perceived behavioural controlhas a significant negative effect on perceived complexity(b ¼ �0.64, p < 0.001), which indicates that individuals who feelthey have the capability and resources to purchase online willperceive online shopping as easier to use, consistent with thefindings of other studies (e.g. Hernandez et al., 2009).

The tenth hypotheses that predicted that perceived relativeadvantages of online travel shopping would positively affect in-tentions to purchase travel onlinewas accepted (b¼ 0.00, p¼ 0.38).However, contrary to what was expected, hypothesis nine, thatestablished a positive relationship between perceived relative ad-vantages and intentions to purchase travel online was not sup-ported (b ¼ 0.00, p ¼ 0.38). Yet, Table 6 shows that perceived

Table 5Structural estimates between second-order constructs and other constructs in themodel.

Path Coefficient t-Statistic

Perceived Behavioural Control / Intentions to Purchase 0.10 4.45***Perceived Behavioural Control / Complexity �0.64 42.17***Perceived Relative Advantage / Attitude 0.26 10.94***Perceived Relative Advantage / Intentions to Purchase 0.00 0.32 ns

***Significant at the 0.001 level based on 5000 bootstrap samples; ns e non-significant.

Fig. 2. PLS analysis results.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 73

relative advantages have a significant total effect on intentions topurchase travel online (b ¼ 0.19, p < 0.001). Further examinationsfollowing Holmbeck (1997) recommendations and Sobel's (1982) Z-statistic proved that attitude totally mediated the relationship be-tween perceived relative advantages and intentions to purchasetravel online. Therefore, the unsupported hypothesis must beanalysed with caution. Not only is the total effect of perceivedrelative advantages on intentions to purchase travel online signif-icant, but also it is one of its most important predictors.

Hypothesis 11, that expected a positive relationship betweenperceived relative advantages and trust was confirmed (b ¼ 0.54,p < 0.001). This was an important finding, since this relationshiphas never been explored in the context of online travel shopping.This means that Internet users trust in online shopping can beincreased by emphasising the perceived relative advantages ofonline travel shopping.

Hypotheses 12 and 13 concerned the influence of perceived riskof online travel shopping on attitude and intentions to purchasetravel online. The latter hypothesis was supported (b ¼ �0.12,p < 0.001), while the former was not (b¼�0.03, p¼ 0.18). Althoughperceived risk does not affect attitude towards online shopping, itmay inhibit individuals from purchasing travel online, since itnegatively influences intentions.

Few studies have addressed trust in the context of online travelshopping and the few that have done so reached different con-clusions. Hypotheses 14 and 15, proposing that trust would be

Table 6Direct, indirect and total effects on intentions to purchase.

Construct Direct Indirect Total t-Statistic

Attitude 0.55 e 0.55 22.12***Communicability 0.01 e 0.01 0.51 nsCompatibility 0.29 0.17 0.46 15.56***Complexity e �0.03 �0.03 2.06*Perceived Behavioural Control 0.10 0.02 0.12 5.20***Perceived Relative Advantages 0.00 0.19 0.19 7.53***Perceived Risk �0.12 �0.02 �0.14 4.74***Trust �0.06 0.16 0.10 3.82***

*Significant at the 0.05 level; *** Significant at the 0.001 level; ns e non-significant.

positively associated with attitude and negatively associated withperceived risk were confirmed (b ¼ 0.11, p < 0.001 and b ¼ �0.72,p < 0.001), supporting the findings of Bign�e et al. (2010) and ofKamarulzaman (2007) that trust influences perceived risk andattitude. However, contrary to these authors that did not find adirect impact of trust on intentions, the results of this study supportthis relationship, accepting hypothesis 16 (b ¼ �0.06, p < 0.05). Itshould however be noted that it has a small impact.

To evaluate the predictive power of the research model, a majoremphasis in PLS analysis is to examine the explained variance (R2)of the endogenous constructs4 (Chin, 2010) that indicate theamount of variance in the construct which is explained by themodel.

As shown in Fig. 2, R2 values range from 0.289 to 0.67, whichindicates that the model has high predictive value and is capable ofexplaining endogenous constructs.

Another approach to assess predictive relevance is to apply thepredictive sample reuse technique developed by Geisser (1975) andStone (1974) , known as the StoneeGeisser's Q2. The Q2 values of theendogenous constructs were calculated in SmartPLS, with theblindfolding procedure, and were all superior to zero (ranging from0.17 to 0.58), indicative of the endogenous constructs' predictiverelevance.

5. Conclusions and implications

From a theoretical perspective, this study has made severaladvances. First, because it examines online travel shopping basedon a holistic approach, integrating several theoretical models andvalidates the integration of these theories in the context of onlinetravel shopping. It confirms attitude and perceived behaviouralcontrol as predictors of intentions as postulated in the TRA(Fishbein & Ajzen, 1975), and TPB (Ajzen, 1991), respectively. The

4 It should be noted that when the repeated indicators approach is used, thevariance of the second-order construct is perfectly explained by its lower compo-nents, therefore R2 ¼ 1.0 (Ringle, Sarstedt, & Straub, 2012). Therefore the R2 of thesecond-order constructs is not reported.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7974

study shows that Rogers' (1995) IDT can be used to explain in-tentions to purchase travel online, since the innovations charac-teristics of relative advantages, compatibility and complexity arevalid predictors of intentions to purchase travel online. The resultsalso indicate that the IDT is superior to TAM, since addingcompatibility increases the explained variance. More, the studydemonstrates that perceived relative advantages and complexity(similar to perceived usefulness and perceived ease of use,respectively, from the TAM) are important predictors of attitudetowards online shopping. By integrating all these theories andadding other relevant constructs, a holistic view was obtained,providing more information than studies with fragmented results.

Considering that research with multidimensional constructsusing PLS path modelling is still limited (Wetzels, Odekerken-Schroder, & van Oppen, 2009), this study fills this gap by usingmultidimensional constructs to operationalise three constructs.Using multidimensional constructs enhance the understanding ofthe overall construct (Law et al., 1998), by providing detail ondifferent facets of the construct (Petter et al., 2007). In particular:

- The empirical results have shown that perceived behaviouralcontrol can be conceptualised as a second-order construct,formed by two distinct dimensions: self-efficacy and control-lability. Although Ajzen (2002) recommends decomposingperceived behavioural control in this manner, no study focus-sing on online travel purchasing intentions had conceptualisedperceived behavioural control in this way.

- This study has shown that perceived relative advantages can beoperationalised as a multidimensional construct, composed byconvenience, time saving, financial advantages, enjoyment andproduct variety. Each dimension represents a facet of the ad-vantages of purchasing travel online, identified in the literaturereview.

Although not a part of the main aim of this research, otherfindings have been made concerning the role of several constructsthat were unexplored or had contradictory results. For instance,only one study (Bign�e et al., 2010) considered perceived behav-ioural control and, contrary to what the TPB posits, it found that itdid not affect intentions to purchase travel online. However, thisstudy's results are consistent with the TPB, since they show thatperceived behavioural control affects intentions to purchase travelonline.

At a timewhen Internet use and online travel shopping aremoreprevalent, factors such as perceived behavioural control orperceived complexity with online travel shopping play a small role.What really matters for Internet users to purchase travel online ishaving a favourable attitude towards online travel shopping andfinding it compatible with their lifestyle. Furthermore, individualsthat perceive the advantages of purchasing travel online, namelytime saving, convenience and financial aspects, will be more likelyto purchase travel online. It is also interesting to note that eventhough trust and security in computer systems are increasing(Bogdanovych, Berger, Simoff, & Sierra, 2006) and online shoppingis nowadays a common practice, perceived risk continues tonegatively affect intentions to purchase travel online.

It is evident that attitude towards online travel shopping is themost relevant determinant of intentions to purchase travel online.Therefore, online travel marketers need to pay close attention tothe factors that contribute to a favourable attitude. This study hasevidenced some of those factors, namely trust, complexity andperceived relative advantages.

Fam, Foscht, and Collins (2004) found that in order to increasetrust in online shopping, providers should offer a warranty ofrefund to consumers, reassure that the information provided will

remain confidential and private, provide formal guarantees of ser-vice and/or products, inform if the travel service is available at thetime of purchase and welcome feedback and comments. Providingexplanations of all the costs involved, offering reliable securitymeasures, no disclosure of credit card details, using the latestencryption technology, explaining how the information collectedwill be used, providing a functional navigation and having a well-designed website are other actions that can increase trust in on-line travel shopping (Austin, Ibeh, & Yee, 2006; Chen, 2006; Kimet al., 2011; Wen, 2010).

Since perceived relative advantages is a significant predictor ofattitude towards online travel shopping, online travel providersneed to emphasise the advantages of purchasing travel online,bearing in mind the advantages that potential buyers most value.The results of the current study have revealed that financial ad-vantages are viewed as a major advantage. Therefore, online travelproviders should guarantee the lowest price and offer otherfinancial advantages such as discounts, coupons and other financialincentives. For example, Intercontinental Hotels Group guaranteethat they have the lowest price, by offering the first night for cus-tomers that find a lower price elsewhere. This is even moreimportant considering that travellers have pointed out the lack ofconfidence that they are getting a good deal as one of the mainreasons why they experience frustration online (PhocusWright,2012).

The results support that convenience and time saving are alsoimportant advantages for travellers to purchase travel online. On-line travel providers should provide procedures that are convenientto travellers such as easy payment features or personalised infor-mation based on past behaviours. They can strive to satisfy theseneeds by, for example, making relevant bits of scattered informa-tion convenient and easy to retrieve (PhocusWright, 2012) andmake suggestions based on customers' past bookings.

The second most important predictor of intentions to purchasetravel online in the model is the Internet user's compatibility. AsVijayasarathy (2004) points out, individuals that find online travelshopping compatible with their lifestyles may be “time starved andconstantly exploring ways to reduce the time to complete varioustasks to manage their busy schedules” (p.757). This is furtherreinforced by the results obtained in perceived relative advantages,as time saving and convenience were found to be significant ad-vantages of purchasing travel online. With hundreds of optionsresulting from an online search, travellers are often overloaded andfeel frustrated (PhocusWright, 2012). Therefore, online travel pro-viders should take advantage of technological advances, by ana-lysing travellers' past behaviour to deliver personalised results andoffer relevant promotions. Furthermore, considering that one thirdof mobile users are planning on the go (Koumelis, 2012), onlinetravel providers should provide apps for mobile devices and tabletsto purchase travel, with other features that facilitate convenience,such as boarding pass or check in.

The results also support the important role that perceived riskplays in purchasing travel online. Therefore, online travel providersmust provide effective ways to reduce users' perceived risk toenhance Internet users' willingness to purchase travel online. Astrong firm reputation is important to reduce the risk associatedwith online shopping (Cases, 2002; Eastlick, Lotz, & Warrington,2006; Kim et al., 2009). Consequently, online travel providersshould strive to build a good reputation, by, for example, cooper-ating with partners who already have a good reputation (Grabner-Kraeuter, 2002), but also by providing fulfilling and satisfyingtransactions. Other risk reduction strategies are to provide travel-lers with information on their consumer rights and personal data,having a security approval symbol (e.g. VeriSign), provide contactinformation, offer money back guarantees, and have high security

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 75

standards that should be communicated to consumers, as well asclear privacy information protecting consumers' personal infor-mation (Bign�e et al., 2010; Cases, 2002; Kim et al., 2009; Lin et al.,2009; Powley et al., 2004; Vijayasarathy, 2004).

Trust presents a statistical significant impact on intentions topurchase travel online. However it is almost insignificant, contra-dicting a recent study that found that trust was the main predictorof intentions to purchase flights from low-cost carrier websites(Escobar-Rodríguez& Carvajal-Trujillo, 2014). This finding points tothe need for further investigation on the role of trust, as it isreasonable to assume that it may vary depending on the type oftravel product being bought.

6. Limitations and future research

Even though the proposed model has been developed on a richtheoretical background, as in any research project, this study haslimitations. First, the data used in this study were based on aconvenience sample and, therefore, generalisations of the resultsmust be made with caution. Although being a strong sample interms of diversity and size, the examination of the respondents'characteristics shows a high proportion of respondents with highereducation levels, which should not be ignored when assessing theconclusions. The replication of this study with a more balancedproportion of Internet users regarding the country of residencewould be desirable.

Another limitation of this study is related to the definition ofonline travel purchases. In this study, the definition considered isbroad, since it includes the purchase of airline tickets, cruises,holiday packages and hotel reservations. Thus, further studiesshould study online purchasing motivations considering distincttravel product categories, rather than considering travel as onecategory. Indeed, a few studies (Bogdanovych et al., 2006;Kamarulzaman, 2007) has found that travellers usually purchaseless complex travel online and prefer booking complex travel froma travel agent. Therefore, the results obtained in this study maydiffer if applied only to low complexity travel or to high complexitytravel. On the other hand, the definition does not consider thedifferent devises people may use to purchase travel online. Giventhe increasing penetration of smartphones, for example, futureresearch could investigate if the results hold independently of thedevise used.

A third limitation of this studywas that it did not consider cross-cultural issues. For example, Heung's (2003) study concluded thatonline travel purchasers were more likely to be from Westerncountries. In a similar vein, Law, Bai, and Leung (2008) found thatAmericans had a higher propensity to purchase travel online thanthe Chinese. Future research should study which factors in the

Construct Indicators

Intentions to Purchase Travel Online INT1 e If you were to purchase travel the ponline would be…(estimation)*INT2 e I expect to purchase travel online in

Attitude ATT1 e Online travel shopping is a good idATT2 e Online travel shopping is a wise idATT3 e I like the idea of purchasing travelATT4 e Purchasing travel online would beATT5 e Purchasing travel online is appealin

Communicability COMM1 e I have heard about people bookCOMM2 e Many friends have purchased trCOMM3 e It is common for people to purc

Self-Efficacy (first-order construct ofPerceived Behavioural Control)

SEF1 e I am proficient in using the InterneSEF2 e I feel confident that I can use the In

proposed model are culture specific and which are cross-cultural.Studying the differences across nations is important in decidingwhich marketing elements may be standardised globally (Pizam,1999).

The datawas analysed using PLS equationmodelling on the totalsample. However, researchers have pointed out that this assumesthat the data are homogeneous, which can be unrealistic (Ringle,Sarstedt, & Schlittgen, 2010). Therefore, future research may usemulti-group analyses to identify if there is heterogeneity orsegment level differences. In SmartPLS this can be done with thefinite mixture PLS (FIMIX-PLS) tool. This approach identifiesdifferent segments and their estimates for the relationships be-tween constructs in the structural model (Ringle, 2006). A betterunderstanding of the factors that influence intentions to purchasetravel online in different segments enhances the possibilities ofdeveloping online marketing strategies that meet the needs of eachsegment.

Given attitude's importance in explaining intentions to purchasetravel online, it is essential to examine the factors affecting attitudeformation, besides the ones considered in this study. For example,prior experience with online travel purchases (e.g. Morosan &Jeong, 2006), perceived playfulness (e.g. Morosan & Jeong, 2008),enjoyment (e.g. Hassanein & Head, 2007) and personal innova-tiveness (e.g. Limayem et al., 2000) have been found to affect atti-tude towards online shopping.

In a similar vein, although the antecedents of intentions topurchase online explained a substantial amount of its variance,theremay be other important factors which have not been includedin the model, representing an opportunity for further research. Forinstance, satisfaction with previous online purchases (Kim et al.,2006) and consumer shopping orientations (Jensen, 2012) mayexplain intentions to purchase travel online.

Finally, there has been strong evidence that supports the linkbetween intended and actual behaviour. Yet, in the context of on-line shopping, this relationship has been largely overlooked. Futureresearch should assert this relationship between intentions topurchase travel online and actual behaviour, since this relationshiphas never been examined in the travel context.

In spite of several limitations, academic researchers, tourismpractitioners and marketers can take advantage of this study tobetter understand the adoption of online travel shopping andconsequently improve online travel distribution strategies. Therecommendations for further investigations also provide re-searchers with challenging directions for future research.

Appendix 1. Construct indicators

References

robability of purchasing Adapted from Grewal, Monroe, and Krishnan(1998) and Teo and Yeong (2003)

the near future (intention). Adapted from Bign�e et al. (2010) andLimayem et al. (2000)

ea. Adapted from Ajzen and Fishbein (1980)ea.online.pleasant.ging travel online many times. Li and Buhalis (2006) and Morrison

et al. (2001)avel online.hase travel online. New Itemt for travel shopping. Adapted from Vijayasarathy (2004)ternet to purchase travel.

(continued on next page)

(continued )

Construct Indicators References

Controllability (first-order constructof Perceived Behavioural Control)

CONT1 e All necessary resources (e.g. computer, internet access, time)for purchasing travel online are accessible to me.

Pavlou and Fygenson (2006)

CONT2 e I have the necessary financial means (e.g. credit card, Paypal)to purchase travel online.

New Item

Trust TRU1 e The chance of having a technical failure in an online transactionis quite small.

Corbitt et al. (2003)

TRU2 e I believe most e-commerce travel web sites will perform to theoutmost of the customers' benefit.TRU3 e I believe online travel sites are trustworthy. Kim et al. (2011)TRU4 e Internet shopping is unreliable. (R) Adapted from Lee and Turban (2001)TRU5 e Internet shopping cannot be trusted, there are too manyuncertainties. (R)

Compatibility COMP1 e Using the internet to purchase travel is compatible with theway I like to shop.

Vijayasarathy (2004)

COMP2 e Using the Internet to purchase travel fits with my lifestyle.Complexity COMP1 e I feel online purchasing procedures are not clear to me. Li and Buhalis (2006)

COMP2 e I feel it is not easy to book travel online.COMP3 e I would find it easy to purchase what I wanted online. (R) Adapted from (Davis, 1989)COMP4 e Purchasing online is easy. (R)

Perceived Risk RISK1 e I do not feel comfortable giving out credit card information tomake a transaction over the Internet.

Cho (2004)

RISK2 e I feel apprehensive about purchasing online. New MeasureRISK3 e Purchasing travel online is risky. Shim et al. (2001)RISK4 e There is too much uncertainty associated with purchasingtravel online.RISK5 e Compared with other methods of purchasing, shoppingonline is riskier.

Convenience (First-order constructof Perceived Relative Advantage)

CONV1 e Purchasing travel online makes me less dependent ofopening hours.

Adapted from Verhoef and Langerak (2001)

CONV2 e Purchasing travel online has easy payment procedures.CONV3 e Purchasing travel online is more convenient thanregular shopping,as I can do it anytime and anywhere.

Adapted from Limayem et al. (2000)

Financial Advantages (First-OrderConstruct of Perceived RelativeAdvantage)

FADV1 e I save money by purchasing travel online. Limayem et al. (2000)FADV2 e Online travel shopping provides more discounts thanoffline travel purchasing.

Kim et al. (2011)

FADV3 e Generally, travel websites offer tourism products atcheaper prices.

Adapted from Li and Buhalis (2006)

Time Saving (First-Order Constructof Perceived Relative Advantage)

TSAV1 e Purchasing travel online enables (will enable) me tocomplete shopping quickly.

Adapted from Davis (1989)

TSAV2 e I can save time by purchasing travel online. Adapted from Limayem et al. (2000)TSAV e Purchasing travel online takes less time than purchasing attravel agencies.

Cho (2004)

Enjoyment (First-Order Constructof Perceived Relative Advantage)

EJY1 e Purchasing travel online is more exciting than purchasingoffline.

Adapted from Verhoef and Langerak (2001)

EJY2 e Purchasing travel online enjoys me more than purchasingoffline.

Childers, Carr, Peck, and Carson (2001)

Product Variety (Dimension ofPerceived Relative Advantage)

PVAR1 e There is a larger choice of travel products availablewhen purchasing online.

Jensen (2009)

PVAR2 e The Internet allows me to purchase travel services thatare not available offline.PVAR3 e I can design a custom made trip by purchasing travel online. New Measure

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7976

Appendix 2. Factor Loadings (bolded) and cross loadings

ATT CMM CMP CXY CTR CNV EJY FAD RSK PVR INT SEF TSV TRT

ATT1 0.90 0.46 0.68 �0.46 0.42 0.55 0.34 0.46 �0.49 0.40 0.78 0.52 0.46 0.54ATT2 0.94 0.44 0.72 �0.46 0.41 0.54 0.39 0.46 �0.49 0.42 0.77 0.52 0.45 0.55ATT3 0.91 0.40 0.67 �0.44 0.35 0.51 0.41 0.45 �0.45 0.40 0.67 0.44 0.47 0.51ATT4 0.90 0.44 0.66 �0.46 0.38 0.54 0.38 0.46 �0.49 0.40 0.69 0.48 0.46 0.55ATT5 0.66 0.39 0.43 �0.31 0.26 0.37 0.30 0.36 �0.29 0.30 0.41 0.29 0.37 0.36CMM1 0.41 0.85 0.27 �0.29 0.31 0.33 0.13 0.23 �0.21 0.18 0.32 0.28 0.29 0.26CMM2 0.44 0.90 0.30 �0.30 0.32 0.34 0.15 0.26 �0.25 0.20 0.36 0.29 0.30 0.29CMM3 0.38 0.77 0.27 �0.26 0.23 0.31 0.18 0.25 �0.21 0.23 0.30 0.24 0.25 0.26CMP1 0.74 0.36 0.94 �0.51 0.44 0.58 0.41 0.46 �0.55 0.43 0.71 0.56 0.45 0.60CMP2 0.63 0.24 0.91 �0.46 0.37 0.52 0.41 0.40 �0.53 0.39 0.58 0.49 0.40 0.57CXY1 �0.29 �0.19 �0.34 0.69 �0.33 �0.32 �0.13 �0.23 0.45 �0.15 �0.30 �0.40 �0.23 �0.36CXY2 �0.39 �0.25 �0.39 0.79 �0.38 �0.39 �0.15 �0.27 0.54 �0.16 �0.38 �0.43 �0.29 �0.47CXY3 �0.48 �0.33 �0.48 0.83 �0.51 �0.52 �0.22 �0.36 0.41 �0.26 �0.44 �0.54 �0.43 �0.47

(continued )

ATT CMM CMP CXY CTR CNV EJY FAD RSK PVR INT SEF TSV TRT

CXY4 �0.30 �0.23 �0.34 0.70 �0.37 �0.38 �0.22 �0.23 0.31 �0.23 �0.23 �0.41 �0.31 �0.36CTR1 0.37 0.32 0.35 �0.46 0.87 0.43 0.12 0.26 �0.29 0.18 0.34 0.56 0.35 0.32CTR2 0.36 0.25 0.41 �0.45 0.83 0.43 0.11 0.25 �0.33 0.17 0.40 0.43 0.28 0.33CNV1 0.43 0.29 0.42 �0.39 0.39 0.79 0.22 0.38 �0.28 0.30 0.38 0.37 0.47 0.35CNV2 0.47 0.32 0.49 �0.50 0.43 0.79 0.32 0.52 �0.42 0.38 0.40 0.39 0.49 0.47CNV3 0.51 0.33 0.52 �0.42 0.39 0.84 0.33 0.48 �0.39 0.41 0.45 0.42 0.54 0.43EJY1 0.36 0.15 0.37 �0.19 0.10 0.30 0.92 0.34 �0.23 0.52 0.27 0.20 0.35 0.31EJY2 0.42 0.18 0.44 �0.25 0.15 0.37 0.94 0.37 �0.28 0.59 0.34 0.24 0.38 0.35FAD1 0.50 0.28 0.47 �0.37 0.31 0.59 0.35 0.89 �0.34 0.47 0.45 0.32 0.49 0.39FAD2 0.44 0.25 0.41 �0.32 0.25 0.50 0.34 0.90 �0.28 0.48 0.39 0.27 0.44 0.35FAD3 0.40 0.23 0.36 �0.28 0.20 0.43 0.32 0.84 �0.26 0.45 0.32 0.22 0.44 0.33RSK1 �0.50 �0.30 �0.50 0.53 �0.31 �0.42 �0.24 �0.34 0.83 �0.25 �0.48 �0.42 �0.33 �0.69RSK2 �0.41 �0.20 �0.49 0.46 �0.33 �0.37 �0.21 �0.24 0.80 �0.20 �0.44 �0.44 �0.24 �0.56RSK3 �0.39 �0.17 �0.49 0.44 �0.28 �0.36 �0.23 �0.27 0.80 �0.21 �0.40 �0.42 �0.26 �0.57RSK4 �0.35 �0.17 �0.36 0.34 �0.22 �0.28 �0.21 �0.22 0.75 �0.20 �0.36 �0.29 �0.22 �0.50RSK5 �0.41 �0.22 �0.45 0.45 �0.28 �0.37 �0.22 �0.25 0.81 �0.24 �0.43 �0.36 �0.27 �0.60PVR1 0.29 0.15 0.28 �0.14 0.10 0.30 0.48 0.39 �0.15 0.81 0.24 0.14 0.33 0.22PVR2 0.33 0.21 0.34 �0.21 0.17 0.37 0.47 0.44 �0.20 0.86 0.29 0.19 0.38 0.27PVR3 0.47 0.23 0.47 �0.31 0.23 0.45 0.55 0.50 �0.32 0.83 0.43 0.26 0.43 0.37INT1 0.75 0.38 0.69 �0.47 0.43 0.51 0.32 0.43 �0.55 0.38 0.94 0.54 0.40 0.54INT2 0.73 0.36 0.63 �0.39 0.38 0.45 0.31 0.41 �0.45 0.36 0.94 0.46 0.37 0.46SEF1 0.44 0.27 0.51 �0.53 0.52 0.43 0.22 0.28 �0.42 0.23 0.46 0.93 0.33 0.42SEF2 0.54 0.32 0.55 �0.57 0.56 0.49 0.22 0.30 �0.48 0.22 0.54 0.94 0.33 0.49TSV1 0.49 0.34 0.44 �0.41 0.36 0.59 0.35 0.49 �0.33 0.41 0.40 0.34 0.91 0.41TSV2 0.49 0.30 0.44 �0.42 0.36 0.60 0.34 0.48 �0.33 0.39 0.39 0.35 0.92 0.40TSV3 0.37 0.25 0.34 �0.29 0.26 0.46 0.35 0.41 �0.23 0.43 0.30 0.23 0.82 0.31TRT1 0.36 0.20 0.39 �0.36 0.23 0.34 0.28 0.25 �0.43 0.24 0.29 0.33 0.29 0.66TRT2 0.37 0.25 0.42 �0.34 0.21 0.35 0.27 0.31 �0.36 0.26 0.28 0.28 0.33 0.65TRT3 0.55 0.30 0.59 �0.45 0.32 0.48 0.34 0.39 �0.59 0.35 0.47 0.41 0.40 0.82TRT4 0.44 0.23 0.47 �0.45 0.33 0.39 0.22 0.29 �0.69 0.22 0.45 0.40 0.28 0.83TRT5 0.48 0.24 0.50 �0.47 0.33 0.40 0.24 0.31 �0.65 0.26 0.48 0.41 0.31 0.81

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 77

References

Ajzen, I. (1985). From intentions to actions: a theory of planned behavior. In J. Kuhl,& J. Beckman (Eds.), Action-control: From cognition to behavior (pp. 11e39).Heidelberg, Germany: Springer.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and HumanDecision Processes, 50(2), 179e211.

Ajzen, I. (2002). Perceived behavioral control, self-efficacy, locus of control, andthe theory of planned behavior. Journal of Applied Social Psychology, 32(4),665e683.

Ajzen, I. (2011). Behavioral interventions: design and evaluation guided by thetheory of planned behavior. In M. M. Mark, S. I. Donaldson, & B. C. Campbell(Eds.), Social psychology for program and policy evaluation (pp. 74e100). NewYork, NY: Guilford.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting socialbehavior. Englewood Cliffs, NJ: Prentice-Hall.

Amaro, S., & Duarte, P. (2013). Online travel purchasing: a literature review. Journalof Travel & Tourism Marketing, 30(8), 755e785.

Andreev, P., Heart, T., Maoz, H., & Pliskin, N. (2009). Validating formative partialleast squares (PLS) models: methodological review and empirical illustration. InPaper presented at the thirtieth international conference on information systems.Phoenix, Arizona.

Armitage, C. J., & Conner, M. (2001). Efficacy of the theory of planned behaviour: ameta-analytic review. British Journal of Social Psychology, 40(4), 471e499.

Austin, N., Ibeh, K., & Yee, J. C. C. (2006). Consumer trust in the online travelmarketplace. Journal of Internet Commerce, 5(2), 21e39.

Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models.Journal of the Academy of Marketing Science, 16(1), 74e94.

Barclay, D., Higgins, C., & Thompson, R. (1995). The partial least squares (PLS)approach to causal modeling: personal computer adoption and use as anillustration. Technology Studies, 2(2), 285e309.

Becker, J.-M., Klein, K., & Wetzels, M. (2012). Hierarchical latent variable models inPLS-SEM: guidelines for using reflective-formative type models. Long RangePlanning, 45(5e6), 359e394.

Bellman, S., Lohse, G. L., & Johnson, E. J. (1999). Predictors of online buying behavior.Communications of the ACM, 42(12), 32e38.

Bign�e, E., Sanz, S., Ruiz, C., & Ald�as, J. (2010). Why some internet users don't buy airtickets online. In U. Gretzel, R. Law, & M. Fuchs (Eds.), Information andcommunication technologies in tourism 2010 (pp. 209e221). Vienna, Austria:Springer.

Bogdanovych, A., Berger, H., Simoff, S., & Sierra, C. (2006). Travel agents vs. onlinebooking: tackling the shortcomings of nowadays online tourism portals. InM. Hitz, M. Sigala, & J. Murphy (Eds.), Information and communication technol-ogies in tourism 2006 (pp. 418e428). Vienna, Austria: Springer.

Buhalis, D. (1998). Strategic use of information technologies in the tourism industry.Tourism Management, 19(5), 409e421.

Buhalis, D., & Law, R. (2008). Progress in information technology and tourismmanagement: 20 years on and 10 years after the Internetdthe state of eTour-ism research. Tourism Management, 29(4), 609e623.

Burnkrant, R. E., & Cousineau, A. (1975). Informational and normative social influ-ence in buyer behavior. Journal of Consumer Research, 2, 206e215.

Cases, A. S. (2002). Perceived risk and risk-reduction strategies in Internet shopping.International Review of Retail, Distribution and Consumer Research, 12(4),375e394.

Chen, C. (2006). Identifying significant factors influencing consumer trust in anonline travel site. Information Technology & Tourism, 8(3/4), 197e214.

Chen, L., Gillenson, M. L., & Sherrell, D. L. (2002). Enticing online consumers: anextended technology acceptance perspective. Information &Management, 39(8),705e719.

Childers, T. L., Carr, C. L., Peck, J., & Carson, S. (2001). Hedonic and utilitarian mo-tivations for online retail shopping behavior. Journal of Retailing, 77(4), 511e535.

Chin, W. W. (1998). Commentary: issues and opinion on structural equationmodeling. MIS Quarterly, 22(1), viiexvi.

Chin, W. W. (2010). How to write up and report PLS analyses. In V. E. Vinzi,W. W. Chin, J. Henseler, & H. Wang (Eds.), Handbook of partial least squares:Concepts, methods and applications in marketing and related fields (pp. 655e690).Berlin: Springer.

Cho, J. (2004). Likelihood to abort an online transaction: influences from cognitiveevaluations, attitudes, and behavioral variables. Information & Management,41(7), 827e838.

Cho, Y. C., & Agrusa, J. (2006). Assessing use acceptance and satisfaction towardonline travel agencies. Information Technology & Tourism, 8(3), 179e195.

Christou, E., & Kassianidis, P. (2003). Consumer's perceptions and adoption of onlinebuying for travel products. Journal of Travel & Tourism Marketing, 12(4), 93e107.

Corbitt, B. J., Thanasankit, T., & Yi, H. (2003). Trust and e-commerce: a study ofconsumer perceptions. Electronic Commerce Research and Applications, 2(3),203e215.

Corritore, C. L., Kracher, B., & Wiedenbeck, S. (2003). On-line trust: concepts,evolving themes, a model. International Journal of Human-Computer Studies,58(6), 737e758.

Davis, F. D. (1985). A technology acceptance model for empirically testing new end-user information systems: Theory and results. Doctoral Dissertation. SloanSchool of Management, Massachusetts Institute of Technology.

Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptanceof information technology. MIS Quarterly, 13(3), 319e340.

Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computertechnology: a comparison of two theoretical models. Management Science,35(8), 982e1003.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e7978

Diamantopoulos, A., & Winklhofer, H. M. (2001). Index construction with formativeindicators: an alternative to scale development. Journal of Marketing Research,269e277.

Eastlick, M., Lotz, S., & Warrington, P. (2006). Understanding online B-to-C re-lationships: an integrated model of privacy concerns, trust, and commitment.Journal of Business Research, 59(8), 877e886.

eMarketer. (2012). Online as a percentage of total travel sales to outrank Asia-Pacificin 2013. Retrieved April 15, 2013, from http://www.emarketer.com/Article/Online-Travel-Sales-Explode-Latin-America/1009493.

Escobar-Rodríguez, T., & Carvajal-Trujillo, E. (2014). Online purchasing tickets forlow cost carriers: an application of the unified theory of acceptance and use oftechnology (UTAUT) model. Tourism Management, 43, 70e88.

Fam, K. S., Foscht, T., & Collins, R. D. (2004). Trust and the online relationshipdanexploratory study from New Zealand. Tourism Management, 25(2), 195e207.

Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behaviour: An intro-duction to theory and research. Reading, MA: Addison-Wesley.

Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with un-observable variables and measurement error. Journal of Marketing Research,18(1), 39e50.

Gefen, D., Rigdon, E. E., & Straub, D. (2011). An update and extension to SEMguidelines for administrative and social science research. MIS Quarterly, 35(2),IIIeXIV.

Geisser, S. (1975). The predictive sample reuse method with applications. Journal ofthe American Statistical Association, 70(350), 320e328.

George, J. F. (2004). The theory of planned behavior and Internet purchasing.Internet Research, 14(3), 198e212.

Grabner-Kraeuter, S. (2002). The role of consumers' trust in online-shopping.Journal of Business Ethics, 39(1/2), 43e50.

Gretzel, U., & Fesenmaier, D. R. (2009). Information technology: shaping the past,present, and future of tourism. In T. Jamal, & M. Robinson (Eds.), The SAGEhandbook of tourism studies (pp. 558e580). London: Sage.

Grewal, D., Monroe, K. B., & Krishnan, R. (April 1998). The effects of price-comparison advertising on buyers' perceptions of acquisition value, trans-action value, and behavioral intentions. Journal of Marketing, 62, 46e59.

Hair, J. F., Black, W. C., Babin, H. J., & Anderson, R. E. (2010).Multivariate data analysis(7th ed.). New Jersey: Prentice Hall.

Hair, J. F., Ringle, C. M., Hult, G. T. M., & Sarstedt, M. (2013). A primer on partial leastsquares structural equation modeling (PLS-SEM). Thousand Oaks, CA: SAGEPublications.

Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011). PLS-SEM: indeed a silver bullet. Journalof Marketing Theory and Practice, 19(2), 139e151.

Hair, J. F., Sarstedt, M., Pieper, T. M., & Ringle, C. M. (2012). The use of partial leastsquares structural equation modeling in strategic management research: a re-view of past practices and recommendations for future applications. Long RangePlanning, 45(5e6), 320e340.

Hair, J. F., Sarstedt, M., Ringle, C. M., & Mena, J. A. (2012). An assessment of the use ofpartial least squares structural equation modeling in marketing research.Journal of the Academy of Marketing Science, 40(3), 414e433.

Hassanein, K., & Head, M. (2007). Manipulating perceived social presence throughthe web interface and its impact on attitude towards online shopping. Inter-national Journal of Human-Computer Studies, 65(8), 689e708.

Henseler, J., Ringle, C. M., & Sarstedt, M. (2012). Using partial least squares pathmodeling in international advertising research: basic concepts and recent is-sues. In S. Okazaki (Ed.), Handbook of research in international advertising (pp.252e276). Cheltenham, UK: Edward Elgar Publishing.

Henseler, J., Ringle, C. M., & Sinkovics, R. R. (2009). The use of partial least squarespath modeling in international marketing. Advances in International Marketing,20, 277e319.

Hernandez, B., Jimenez, J., & Martín, M. J. (2009). Adoption vs acceptance of e-commerce: two different decisions. European Journal of Marketing, 43(9/10),1232e1245.

Heung, V. C. S. (2003). Internet usage by international travellers: reasons and bar-riers. International Journal of Contemporary Hospitality Management, 15(7),370e378.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity inthe study of mediators and moderators: examples from the child-clinical andpediatric psychology literatures. Journal of Consulting and Clinical Psychology,65(4), 599e610.

Hsu, M.-H., & Chiu, C.-M. (2004). Predicting electronic service continuance with adecomposed theory of planned behaviour. Behaviour & Information Technology,23(5), 359e373.

Hung, Y.-C., Yang, Y.-L., Yang, H.-E., & Chuang, Y.-H. (2011). Factors affecting theadoption of e-commerce for the tourism industry in Taiwan. Asia Pacific Journalof Tourism Research, 16(1), 105e119.

Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of constructindicators and measurement model misspecification in marketing and con-sumer research. Journal of Consumer Research, 30(2), 199e218.

Jensen, J. M. (2009). Travellers' intentions to purchase travel products online: therole of shopping orientation. In A. Matias, P. Nijkamp, & M. Sarmento (Eds.),Advances in tourism economics: New developments (pp. 203e215). Heidelberg,Germany: Physica Verlag.

Jensen, J. M. (2012). Shopping orientation and online travel shopping: the role oftravel experience. International Journal of Tourism Research, 14(1), 56e70.

Kamarulzaman, Y. (2007). Adoption of travel e-shopping in the UK. InternationalJournal of Retail & Distribution Management, 35(9), 703e719.

Karahanna, E., Ahuja, M., Srite, M., & Galvin, J. (2002). Individual differences andrelative advantage: the case of GSS. Decision Support Systems, 32(4), 327e341.

Kim, W. G., & Kim, D. J. (2004). Factors affecting online hotel reservation intentionbetween online and non-online customers. International Journal of HospitalityManagement, 23(4), 381e395.

Kim, D. J., Kim, W. G., & Han, J. S. (2007). A perceptual mapping of online travelagencies and preference attributes. Tourism Management, 28(2), 591e603.

Kim, L. H., Kim, D. J., & Leong, J. K. (2005). The effect of perceived risk on purchaseintention in purchasing airline tickets online. Journal of Hospitality & LeisureMarketing, 13(2), 33e53.

Kim, L. H., Qu, H., & Kim, D. J. (2009). A study of perceived risk and risk reduction ofpurchasing air-tickets online. Journal of Travel & Tourism Marketing, 26(3),203e224.

Kim, M.-J., Chung, N., & Lee, C.-K. (2011). The effect of perceived trust on electroniccommerce: shopping online for tourism products and services in South Korea.Tourism Management, 32(2), 256e265.

Kim, W. G., Ma, X., & Kim, D. J. (2006). Determinants of Chinese hotel customers' e-satisfaction and purchase intentions. Tourism Management, 27(5), 890e900.

Kolsaker, A., Lee-Kelley, L., & Choy, P. C. (2004). The reluctant Hong Kong consumer:purchasing travel online. International Journal of Consumer Studies, 28(3),295e304.

Koumelis, T. (2012). Travelers are confident and frequent mobile shoppers.Retrieved April 26, 2013, from http://www.traveldailynews.com/news/article/52052/travelers-are-confident-and-frequent.

Krueger, N. F., Reilly, M. D., & Carsrud, A. L. (2000). Competing models of entre-preneurial intentions. Journal of Business Venturing, 15(5e6), 411e432.

Law, K. S., Chi-Sum, W., & Mobley, W. M. (1998). Toward a taxonomy of multidi-mensional constructs [Article] Academy of Management Review, 23(4), 741e755.

Law, R., Bai, B., & Leung, B. (2008). Travel website uses and cultural influence: acomparison between American and Chinese travelers. Information Technology &Tourism, 10(3), 215e225.

Law, R., Leung, R., & Buhalis, D. (2009). Information technology applications inhospitality and tourism: a review of publications from 2005 to 2007. Journal ofTravel & Tourism Marketing, 26(5e6), 599e623.

Law, R., Leung, K., & Wong, J. (2004). The impact of the internet on travel agencies.International Journal of Contemporary Hospitality Management, 16(2), 100e107.

Lee, Y., & Kozar, K. A. (2008). An empirical investigation of anti-spyware softwareadoption: a multitheoretical perspective. Information & Management, 45(2),109e119.

Lee, J., & Morrison, A. M. (2010). A comparative study of web site performance.Journal of Hospitality and Tourism Technology, 1(1), 50e67.

Lee, M. K. O., & Turban, E. (2001). A trust model for consumer internet shopping.International Journal of Electronic Commerce, 6(1), 75e91.

Lee, H. Y., Qu, H., & Kim, Y. S. (2007). A study of the impact of personal innova-tiveness on online travel shopping behaviorda case study of Korean travelers.Tourism Management, 28(3), 886e897.

Li, L., & Buhalis, D. (2006). E-commerce in China: the case of travel. InternationalJournal of Information Management, 26(2), 153e166.

Limayem, M., Khalifa, M., & Frini, A. (2000). What makes consumers buy frominternet? A longitudinal study of online shopping. IEEE Transactions on Systems,Man, and Cybernetics e Part A: Systems and Humans, 30(4), 421e432.

Lin, H.-F. (2007). Predicting consumer intentions to shop online: an empirical test ofcompeting theories. Electronic Commerce Research and Applications, 6(4),433e442.

Lin, P.-J., Jones, E., & Westwood, S. (2009). Perceived risk and risk-relievers in onlinetravel purchase intentions. Journal of Hospitality Marketing & Management,18(8), 782e810.

McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). Developing and validatingtrust measures for e-commerce: an integrative typology. Information SystemsResearch, 13(3), 334e359.

Mayr, T., & Zins, A. H. (2009). Acceptance of online vs. traditional travel agencies.Anatolia: An International Journal of Tourism and Hospitality Research, 20(1),164e177.

Moital, M., Vaughan, R., & Edwards, J. (2009). Using involvement for segmenting theadoption of e-commerce in travel. The Service Industries Journal, 29(5), 723e739.

Moital, M., Vaughan, R., Edwards, J., & Peres, R. (2009). Determinants of intention topurchase over the internet. Anatolia: An International Journal of Tourism andHospitality Research, 20(2), 345e358.

Moore, G. C., & Benbasat, I. (1991). Development of an instrument to measure theperceptions of adopting an information technology innovation. InformationSystems Research, 2(3), 192e222.

Morosan, C., & Jeong, M. (2006). Understanding travelers' adoption of hotel reser-vation web sites. In M. Hitz, M. Sigala, & J. Murphy (Eds.), Information andcommunication technologies in tourism 2006 (pp. 394e405). Vienna, Austria:Springer.

Morosan, C., & Jeong, M. (2008). Users' perceptions of two types of hotel reservationweb sites. International Journal of Hospitality Management, 27(2), 284e292.

Morrison, A. M., Jing, S., O'Leary, J. T., & Cai, L. A. (2001). Predicting usage of theinternet for travel bookings: an exploratory study. Information Technology &Tourism, 4(1), 15e30.

O'Connor, P., & Murphy, J. (2004). Research on information technology in the hospi-tality industry. International Journal of Hospitality Management, 23(5), 473e484.

Pavlou, P. A., & Fygenson, M. (2006). Understanding and predicting electroniccommerce adoption: an extension of the theory of planned behavior. Man-agement Information Systems Quarterly, 30(1), 115.

S. Amaro, P. Duarte / Tourism Management 46 (2015) 64e79 79

Petter, S., Straub, D., & Rai, A. (2007). Specifying formative constructs in informationsystems research. MIS Quarterly, 31(4), 623e656.

PhocusWright. (2012). Empowering inspiration: The future of travel search. Sherman,CT: PhocusWright.

Pizam, A. (1999). Cross cultural tourist behavior. In A. Pizam, & Y. Mansfeld (Eds.),Consumer behavior in travel and tourism (pp. 393e408). New York: Routledge.

Powley, J. H., Cobanoglu, C., & Cummings, P. R. (2004). Determinants of online travelpurchases from third-party travel web sites. In A. J. Frew (Ed.), Information andcommunication technologies in tourism 2004 (pp. 424e436). Vienna, Austria:Springer.

Riemenschneider, C. K., Hardgrave, B. C., & Davis, F. D. (2002). Explaining softwaredeveloper acceptance of methodologies: a comparison of five theoreticalmodels. IEEE Transactions on Software Engineering, 28(12), 1135e1145.

Ringle, C. (2006). Segmentation for path models and unobserved heterogeneity: thefinite mixture partial least squares approach. Research Papers on Marketing andRetailing, 35, 1e29.

Ringle, C. M., Sarstedt, M., & Schlittgen, R. (2010). Finite mixture and genetic al-gorithm segmentation in partial least squares path modeling: identification ofmultiple segments in complex path models. In A. Fink, B. Lausen, W. Seidel, &A. Ultsch (Eds.), Advances in data analysis, data handling and business intelli-gence. Berlin: Springer-Verlag.

Ringle, C. M., Sarstedt, M., & Straub, D. W. (2012). A critical look at the use of PLS-SEM in MIS Quarterly. MIS Quarterly, 36(1), iiexiv.

Ringle, C. M., Wende, S., & Will, S. (2005). SmartPLS 2.0 (M3) (version beta). [Soft-ware]. Hamburg, Retrieved from http://www.smartpls.de.

Rogers, E. M. (1995). Diffusion of innovations (4th ed.). New York, NY: The Free Press.San Martín, H., & Herrero, A. (2012). Influence of the user's psychological factors on

the online purchase intention in rural tourism: integrating innovativeness tothe UTAUT framework. Tourism Management, 33(2), 341e350.

Schonlau, M., Fricker, R. D., & Elliott, M. N. (2002). Conducting research surveys via e-mail and the web. Santa Monica, CA: RAND Corporation.

Shih, Y. Y., & Fang, K. (2004). The use of a decomposed theory of planned behavior tostudy Internet banking in Taiwan. Internet Research, 14(3), 213e223.

Shim, S., Eastlick, M. A., Lotz, S. L., & Warrington, P. (2001). An online prepurchaseintentions model: the role of intention to search. Journal of Retailing, 77(3),397e416.

Shin, D. H. (2010). MVNO services: policy implications for promoting MVNOdiffusion. Telecommunications Policy, 34(10), 616e632.

Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structuralequation models. Sociological Methodology, 13(1), 290e312.

Stone, M. (1974). Cross-validatory choice and assessment of statistical predictions.Journal of Royal Statistical Society, Series B, 36(2), 111e147.

Suh, B., & Han, I. (2002). Effect of trust on customer acceptance of Internet banking.Electronic Commerce Research and Applications, 1(3e4), 247e263.

Teo, T. S. H., & Yeong, Y. D. (2003). Assessing the consumer decision process in thedigital marketplace. International Journal of Management Science, 31(5),349e363.

Terry, D. (1993). Self-efficacy expectancies and the theory of reasoned action. InD. J. Terry, C. Gallois, & M. McCamish (Eds.), The theory of reasoned action: Itsapplication to AIDS-preventive behaviour (Vol. 28); (pp. 135e151). Queensland,Australia: Pergamon Press.

Tornatzky, L. G., & Klein, K. J. (1982). Innovation characteristics and innovationadoption-implementation: a meta-analysis of findings. IEEE Transactions onEngineering Management, 29(1), 28e45.

Trafimow, D., Sheeran, P., Conner, M., & Finlay, K. A. (2002). Evidence that perceivedbehavioural control is a multidimensional construct: perceived control andperceived difficulty. British Journal of Social Psychology, 41(1), 101e121.

Urbach, N., & Ahlemann, F. (2010). Structural equation modeling in informationsystems research using partial least squares. Journal of Information TechnologyTheory and Application, 11(2), 5e40.

Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technologyacceptance model: four longitudinal field studies. Management Science, 46(2),186e204.

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance ofinformation technology: toward a unified view. MIS Quarterly, 425e478.

Verhoef, P. C., & Langerak, F. (2001). Possible determinants of consumers' adoptionof electronic grocery shopping in the Netherlands. Journal of Retailing andConsumer Services, 8(5), 275e285.

Vijayasarathy, L. R. (2004). Predicting consumer intentions to use on-line shopping:the case for an augmented technology acceptance model. Information & Man-agement, 41(6), 747e762.

Wang, Y. D., & Emurian, H. H. (2005). An overview of online trust: concepts, ele-ments, and implications. Computers in Human Behavior, 21, 105e125.

Wang, M. S., Chen, C. C., Chang, S. C., & Yang, Y. H. (2007). Effects of online shoppingattitudes, subjective norms and control beliefs on online shopping intentions: atest of the theory of planned behaviour. International Journal of Management,24(2), 296e302.

Wen, I. (2010). Online travelers' decision makings: a new equation model to eval-uate impacts of website, search intention, and trust. Information Technology &Tourism, 12(2), 153e173.

Wetzels, M., Odekerken-Schroder, G., & van Oppen, C. (2009). Using PLS pathmodeling for assessing hierarchical construct models: guidelines and empiricalillustration. MIS Quarterly, 33(1), 177e195.

Wong, J., & Law, R. (2005). Analysing the intention to purchase on hotel websites: astudy of travellers to Hong Kong. International Journal of Hospitality Manage-ment, 24(3), 311e329.

Wu, J.-H., & Wang, S.-C. (2005). What drives mobile commerce? An empiricalevaluation of the revised technology acceptance model. Information & Man-agement, 42(5), 719e729.

Yu, T.-K., & Wu, G.-S. (2007). Determinants of internet shopping behavior: anapplication of reasoned behaviour theory. International Journal of Management,24(4), 744e762.

Zhou, Z. (2004). E-commerce and information technology in hospitality and tourism.Canada: Delmar Learning.

Suzanne Amaro has a PhD in Marketing and Strategy, anMSc degree in Management and a five-year bachelor's de-gree in Economics. She started her professional activity asa consultant at a Business and Innovation Centre. She hasbeen an associate professor at the Management Depart-ment of the Polytechnic Institute of Viseu since 1999 andis now head of the Marketing BSc Degree. Her currentresearch interests include: travellers' online purchasebehaviour, the use of social media for travel purposesand social media marketing. She also is very interestedin the use of partial least squares structural equationmodelling as a research technique.

Paulo Duarte is professor of marketing at the Business andEconomics Department and head of the Master in Mar-keting at University of Beira Interior, Portugal. Prior toreceiving his Ph.D. in Management at the University ofBeira Interior, he held a senior marketing position in a fastmoving consumer goods distribution company. Academi-cally, he has been doing research in the fields of consumerbehaviour, satisfaction, brand management, and webmarketing, having published articles on these topics. He isalso member of the editorial board of several internationaljournals.