Social media in Tourism: Establishing factors influencing attitudes towards the usage of Social Networking Sites for trip organisation

Significance of work: The Theory of Planned Behaviour by Ajzen (1988) specifies that behaviour intention is influenced by attitude, subjective norm and perceived behavioural control. The current study extended the Theory of Planned Behaviour by adding two new variables, perceived usefulness and perceived risk to test their effect on attitude. It was concluded that perceived usefulness and perceived risk play a significant role in determining an individual’s attitude towards the use of social networking sites for trip organisation.


INTRODUCTION
Over the last decade, the growth of Internet technologies, especially social media platforms such as social networking sites (SNS) (Facebook, Twitter, MySpace vibe, Pinterest, Instargarm) has increased considerably. Bhakuni and Aronkar (2012) as well as Richard and Guppy (2014) observed that the usage of social Networking Sites (hereafter referred to as SNS) grew rapidly from a platform which serves a few people online, into a platform that is used by a significant number of Internet users. These sites have evolved from a basic online tool for content sharing to become an important part of the media landscape (Singh, Lehnert & Bostick, 2012:685). Statistics shows there were 2.078 billion social media accounts which were active in January 2015 globally, with Facebook leading with a total of 1, 4 billion users (Kemp, 2015:2;Statista, 2015:1). In South Africa, 11.8 million people (22% of the total population) are on Facebook and the number of people on YouTube and Instagram has increased by 53% and 65% respectively between August 2014 and August 2015 (South Africa Social Media Landscape, 2015:3). Gong (2012:421) pointed out that among the Internet-related technologies, social networking sites are the fastest growing.
Unique characteristics of SNS including their interactive nature, the ability to show videos and pictures, screening and filtering of information by friends, have enabled them to attract large numbers of users across the globe (Lange-Faria & Elliot, 2012). The penetration of mobile devices, particularly smartphones, which provide travellers with the means to connect on-the-go and interact in real time, also made social networking more attractive than any other communication tools (Gong, 2012:422). A total of 8.8 million South Africans primarily accessed Facebook in 2015 through mobile phones such as Android, Blackberry and Windows (South Africa Social Media Landscape, 2015:3). The fact that travel information can now be easily accessed through mobile phones and computers has special implications for organisations operating in the tourism industry. A study by Lyu and Wang (2015) revealed that due to the introduction of new Internet technologies (e.g. social media), the number of individuals who use traditional information centres (travel agents and tour operators) to obtain travel information, has dropped by 27.6% in Korea. They noted that travellers now prefer social media to traditional sources of information (Lyu & Wang, 2015). It is therefore important for tourism organisations and tourism policy makers to understand the different factors that influence travellers' use of social networking sites when organising a trip. Establishing these factors will go a long way in assisting tourism managers and policy makers to gain knowledge on how best to appeal to users of their social networking sites. Currently, TripAdvisor, the leader among travel-related consumer reviews on a social networking site (TripAdvisor, 2015), serves more than 200 million people and the application is being downloaded at a rate of 28 times per minute (TripAdvisor, 2015). TripAdvisor is therefore playing a major role in trip organisation when using search engines (Di Pietro & Di Virgilio, 2012: 62;Greenleigh, 2012;Xiang & Gretzel, 2010:181).
Against this background provided, this study aims at contributing to the existing literature on the attitude and perceptions of individuals towards the usage of SNS for trip organisation purposes. The Theory of Planned Behaviour (TPB) model is applied to the study to help explain intentions to use SNS for trip organisation among travellers in Gauteng, South Africa, making the focus of the study novel as most studies on social media are based on samples drawn from western and Asian countries. The current study attempts to extend the TPB model by adding another factor which is perceived risk. A combination of TPB variables with perceived risk may provide a model which captures extensive elements which can better explain SNS adoption behaviour by customers in the tourism industry in South Africa. The TPB model is considered suitable for the study due to the fact that it is regarded as one of the most effective models in predicting online technology adoption behaviour amongst consumers (Jalilvand & Samiei, 2012:593;Quintal, Lee & Soutar, 2010:798;Troung, 2009:179;Hsu, Yen, Chiu & Chang, 2006:890). Therefore, checking the applicability of TPB model in explaining SNS usage from a travel organising perspective (Lopez, Bulchand-Gidumal, Tano & Armas, 2011:642), could assist tourism organisation managers and destination policy makers to have an insight of the factors which affect user acceptance. The study addresses the overall use of social media for trip organisation and is not restricted to a specific social networking site.

PROBLEM STATEMENT
South African Tourism has a digital marketing partnership with social networking sites such as YouTube, TripAdvisor and Facebook (South African Tourism Review, 2015). For example, in 2013, the Cape Town Tourism Board won the Social Media in Travel and Tourism (SMITTY) award for the innovative use of online technologies for marketing Cape Town tourism attractions (Cape Town Tourism, 2013). This illustrates that the South African tourism industry is using social networking sites to attract visitors. It is therefore important to determine how individual travellers are utilising social networking sites when organising trips in order to improve on the effective utilisation of social networking sites. More importantly, tourism organisations have to be aware of the different factors that influence individual traveller use of SNS before, during and after the trip (Lopez et al., 2011:643). Such an understanding remains unclear and therefore requires further investigation.

MAIN OBJECTIVE
To determine the factors that influence customers' attitude towards social networking sites for trip organisation and the influence of attitude on use intentions.

Secondary objectives
 To examine factors that influence the attitude towards the use of social networking sites for trip organisation.  To establish the relative power of factors that influence the attitude towards social networking sites.  To establish whether attitude towards social networking sites has a direct influence on behavioural intentions to use and to recommend the use of SNS for trip organisation.  To propose and test a conceptual model on precursors of attitude and behavioural intentions towards use of SNS for trip organisation.

Theoretical background
Theory of Reasoned Action (TRA) was formed in a bid to determine factors which influences attitude and behaviour (Fishbein & Ajzen, 1975). TRA is centred on the aspect that subjective norm and attitude towards behaviour are the two factors which affect behavioural intention (Fishbein & Ajzen, 1975:302). TRA has been successfully tested by a number of studies (Abadi & Nematizadeh, 2012;Peslak, Ceccucci & Sendall, 2011;Pelling & White, 2009;Porter & Donthu, 2006) to predict behavioural intention in online technology acceptance. However, despite the applicability of TRA in predicting online technology acceptance in so many fields, the model had some limitations which were identified by Ajzen (1988). The major limitation of TRA is that it can only be applied successfully if the behaviour in question is under the will of the person. If behaviour is not under the person's will and control, that individual may not perform the behaviour due to other environmental conditions that might intervene (Ajzen, 1988). This has led to the formulation of the Theory of Planned Behaviour.

Theory of planned behaviour
Theory of Planned Behaviour (TPB) was proposed by Ajzen (1988), it has an additional factor to those in TRA, perceived behavioural control. The theory is centred on three kinds of beliefs (attitude, subjective norm and perceived behavioural control). Perceived behavioural control means the ability to perform certain behaviour and it influences the intention to perform the behaviour (Zoonen, Verhoeven & Elving, 2014:166). According to the TPB model, when a person has a positive attitude towards a behaviour and the influence of friends is higher, the individual's conviction that he/she can perform the behaviour increases, thereby strengthening the intention to perform the behaviour as well (Ajzen, 1991:187). According to the model, one aspect which is on the centre of human behaviour is intention to perform and this aspect is influenced by attitude towards behaviour, the amount of pressure received from friends, and one's perceived ability to perform the behaviour . Thus, in this study it is postulated that, variables in the TPB model, i.e. perceived behavioural control and subjective norm significantly influence the attitude towards the use of SNS with regard to travelling.

SOCIAL NETWORKING SITES AND TRIP ORGANISATION
Social networking sites are defined by Hoffman and Novak (2012:1) "as web-based applications that permit creation, sharing, manipulation and consumption of user-generated content". They provide individuals with platforms to create an identity online and present their image (Hollenbeck & Kaikati, 2012:396). There are currently a number of social networking sites used by people to chat and share information about their trips, something which was not possible just over a decade ago (Senthil, Prabhu & Bhuvaneswari, 2013:51). The first social networking site according to Boyd and Ellison (2007:211), was introduced in 1997, Six degree.com, which allowed users to create a profile and list their friends. This was followed by Friendster.com launched in 2002. After this, there was a wave of social networking sites with different focus e.g. LinkedIn (2003) for business, and Myspace (2003), Facebook (2005) and Twitter (2006) for general discussions. WeChat, Pinterest and Instagram are some of the social networking sites that have gained momentum in the last 3 years. As indicated by Gong (2012:422), users of these social networking sites are not passive content consumers, but also active content generators and distributors. Given their increasing popularity, social networking sites have significantly impacted the way people consume information, socialise, and search for travel information as well as organising trips.
As the uptake of social networking sites increases, academic research surrounding the usage of social networking sites in tourism is also growing. Di Pietro and Di Virgilio (2012) for example, studied the usage of social networking sites on destination choice and discovered that most tourists use social networking sites extensively for choosing holiday accommodation. Xiang and Gretzel (2010:184) indicate that social networking sites constitute a considerable part of online tourist domain and play a vital role when people are organising trips. In line with the TPB, this study investigates the influence or perceived usefulness, subjective norm and perceived behavioural control on attitude towards the use of SNS for trip organisation. It also looks at the influence of perceived risk on attitude, as well as the influence of attitude on behavioural intentions.

ATTITUDE, ITS PRECURSORS AND OUTCOMES
Attitude towards behaviour is described "as the extent to which an individual has a favourable or unfavourable assessment of the behaviour in question" (Ajzen, 1991:188). If the attitude towards behaviour is positive, the individual's desire to perform that particular behaviour increases (Ajzen, 1991:188). Attitudes can be described as a person's overall assessment of performing a particular behaviour (Celik & Yilmaz, 2011:158). Section 3.1 discusses factors that may help explain attitude towards the use of SNS for trip organisation.

Precursors of attitude Perceived usefulness (PU)
Perceived usefulness is defined as "the extent to which a person is convinced that using a particular technology would improve their performance" (Henderson & Divett, 2003:394). In the case of this study, perceived usefulness was viewed as the benefits derived from using social networking sites for trip organisation. These benefits according to Lopez et al. (2011), are different and constantly change such that it is not easy to establish them due to the heterogeneous nature of travellers and their ability to use these social networking sites. Lopez et al. (2011:643) categorise these benefits into three dimensions which are functional benefits, social benefits and hedonic benefits. In their study, Lopez et al. (2011) concluded that functional and social benefits significantly influence attitude towards the usage of social media when organising trips. For this study, two types of benefits, functional benefits and social benefits are assessed, following what has been proposed in e-commerce literature (Lopez et al., 2011;Chung & Buhalis, 2008;Gretzel & Yoo, 2008;Jeong, 2008). Chung and Buhalis (2008) and Sigala (2010) opine that functional and social benefits play an important role in determining the use of social networking sites. It is posited in this study that social networking sites perceived that usefulness which is divided into functional and social benefits, can influence the attitude towards the usage of social networking sites for trip organisation. To ascertain this, the following hypothesis was formulated: H1: Perceived usefulness positively influences the attitude towards the use of social networking sites for trip organisation.

Subjective Norm (SN)
Subjective norm is described "as the perceived social pressure to perform or not to perform behaviour" (Ajzen, 1991:188). TPB views influence or pressure from friends and social groups to be important when one is highly motivated to comply with the exerted pressure (Zoonen et al., 2014:166). The effect of subjective norm on behavioural intention has been supported by some previous studies. Zhou (2011) concluded that subjective norm impacts on online community users' participation intention, and Akman (2014) concluded that subjective norm has an effect on behavioural intention to use social media. Lopez-Nicolas, Molina-Castillo and Bouwman (2008) also found the effect of subjective norm on users' intention to play online games. Taking cognisance of the argument in TPB that attitude is related to behavioural intention, this study thus hypothesised that: H2: Subjective norm positively influences the attitude towards the use of social networking sites for trip organisation.

Perceived risk (PR)
Perceived risk has been defined by Bauer (1967:191) as "a combination of the uncertainty and seriousness of the outcome involved". In support of this, Peter and Ryan (1976:185) defined perceived risk "as the expectation of losses associated with purchase and acts as deterrents to purchase behaviour". Perceived risk as proposed by Cunningham (1976), can be decomposed into subfacets a) performance risk and b) psychosocial risk. They further decomposed perceived risk into six categories which are: a) performance, b) financial, c) safety/privacy, d) social, e) time, and d) psychological loss.
In 1971, Roselius classified perceived risk into five groups which are, time loss, psychological loss, financial loss, physical loss and performance loss. Physical risk and financial risk were not included in this study as they are deemed not applicable to social networking sites. Earlier studies (Mannuka & Jarvi, 2014;Skarmeas & Robson, 2008) found that perceived risk affects how consumers perceive value, and thus negatively affects perceived benefits. Perceived risk also reduces the expected benefits of a particular outcome of a decision-making process and consequently reduces the behavioural intention (Skarmeas & Robson, 2008:180). Thus in this study, perceived risk comprises of a) social risk, b) privacy risk, and c) time risk.
Social risk is defined in this study as the potential loss in one's set of friends as a result of using social networking sites e.g. not fitting in the set of your friends. Time risk is defined as the time wasted through searching and learning how to use social networks, which could be spent productively when using other means of trip organisation. "Privacy risk refers to the potential loss of control of personal information", for example if an individual's personal information is used without their knowledge (Mannuka & Jarvi, 2014:223). This study posits that perceived risk has negative effects on perceived usefulness and attitude towards social networking sites. The following hypotheses were thus formulated: H3: Perceived risk has a negative influence on the perceived usefulness of social networking sites for trip organisation.
H4: Perceived risk negatively influences the attitude towards the use of social networking sites for trip organisation.

Perceived behavioural control (PBC)
PBC is essentially the same idea as self-efficacy; the measure of a person's perception of their ability to complete a duty (Yzer, 2012:103), in this case the ability to use social networking sites to organise a trip. PBC, according to Yzer (2012:103), incorporates two specific aspects, which are perceived ability (which is the extent to which one is convinced that they can carry out a task or perform the behaviour), and perceived autonomy (the extent to which one is convinced that he/she can control the actual behaviour). This study looks at PBC in the context of confidence and certainty of being able to use social networking sites when organising trips. Studies by Baker and White (2010), Hocevar, Flanagin and Metzeger (2014), as well as Leng et al. (2011) have found that perceived behavioural control (selfefficacy) exerts significant positive influence attitude towards the use of social networking sites. Accordingly, it is hypothesised in this study that: H5: Perceived behavioural control positively influences the attitude towards the use of social networking sites for trip organisation.

Behavioural intentions and its precursors
Behavioural intention describes those factors that motivate someone to perform certain behaviour and indicates one's desire to try, or how they seriously attempt to perform that particular task (Ajzen, 1991:181). In this study, intention to use defines the willingness to use social networking sites for trip organisation, and intention to recommend refers to the objective to advise others to use social networks when organising trips.

Perceived behavioural control
The effect of perceived behavioural control on intention has been examined in various studies. A study by Alam and Sayuti (2011) found that perceived behavioural control influences the intention to purchase halal food. Martin, Ramamonjiarivelo and Martin (2011) found that perceived behavioural control is an important factor that influences travel intention. However, Sentosa and Nik Mat (2012) examined the applicability of TPB in Internet purchase behaviour and concluded that perceived behavioural control is not an important determinant of behavioural intention. It is assumed in this study that perceived behavioural control can affect one's intention to use social networking sites when organising trips. The following hypothesis is therefore put forward: H6: Perceived behavioural control has a significant effect on the intention to use social networking sites for trip organisation.
Attitude Rauniar, Rawski, Yang and Johnson (2014) examined the applicability of Technology Acceptance Model (TAM) in explaining social media use and concluded that attitude affects the intention to use social networking sites. For the purpose of this study, TAM is defined as a 'causal model which explains factors that affect behavioural intentions in the usage of new information technologies' (Davis, 1989:323). Wei, Lin, Lu and Chuang (2015) examined the intention of users to continue using social networking sites and concluded that attitude can positively influence the intention to continue using a social networking site. Dennis, Merrilees, Jayawardhena and Wright (2009) formulated a conceptual model to explain e-consumer behaviour and concluded that purchase intention is positively influenced by positive attitude. It is assumed in this study that attitude has a positive impact not only on intentions to use, but also on intentions to recommend others to use social networking sites. The following hypotheses are formulated to ascertain the influence of attitude on the intention to use social networking sites and the intention to recommend others to use social networking sites.
H7: Attitude towards the use of social networking sites has a significant positive impact on the intention to use social networking sites for trip organisation.

Research design and sampling
The study used an exploratory research design that was quantitative in nature. The objective of exploratory research is to determine important or key variables and it also considers the nature of certain relationships (Zikmund, 1984:6). This study seeks to establish key factors which influence attitude towards the use of social networking sites, hence the exploratory research design was chosen. A cross-sectional approach was used to collect data by means of questionnaires where data was collected from the sample on one occasion only. According to Bhattacherjee (2012:18), cross-sectional design can be used to provide data for an exploratory enquiry. The target population for this study was all travellers in the city of Johannesburg in South Africa who are aware of one or more travel-related social networking sites and those who uses social networking sites for trip arrangements. The sample was heterogeneous and consisted of both users and non-users of travel-related social networking sites. The study focused on both business travellers and leisure travellers. Respondents who met this criterion were selected to complete the questionnaire. In order to select the sample for this study, a convenience sampling technique was used where only those individuals who could be easily accessed were asked to answer the questionnaire. The convenient sampling technique involves the selection of the most accessible subjects (Marshal, 1996:523). The questionnaire was administered by trained fieldworkers. A total of 340 questionnaires were returned and used for analysis.

Measurement development
To ensure validity of measurement items used for this study items were borrowed from existing measures (Quintal et al., 2010).The questionnaire items were adopted from existing literature (see Appendix A), but adapted to fit the use of social networking sites for trip organisation. The study used a five-point Likert scale to ensure validity since the studies in which the questionnaire items were adopted also used a five-point Likert scale. The scales were ranging from strongly disagree to strongly agree, except for those measures which solicited demographic information. The survey instrument was pretested on 20 travelers (users and non-users who are aware of any travel social networking sites). The pretesting was done mainly to check the ease of understanding of the questions by respondents. After the pretests, the questions were so that all respondents can understand the requirements of the questionnaire.
For the analysis of data, SPSS Version 21 was used. The reliability and validity of the constructs were tested using Confirmatory Factor Analysis (CFA). Structural Equation Modelling through the use of Partial Least Squares was used for hypotheses testing. Table 1 provides the background information of the respondents. The other social networking sites indicated by respondents were Google, followed by Pinterest.

Analysis and results
The PLS technique was used because it is suitable for examining compound relationships for example, where there are large numbers of variables by avoiding inadmissible solutions and factor indeterminacy (Chin, 1998:8). The PLS technique also allows the testing of hypotheses simultaneously, even if there are measures with single and multiple constructs (Fornell & Bookstein, 1982:43).

Measurement model
The model included 36 items describing 10 latent constructs: attitude, perceived behavioural control, functional benefits, social benefits, time risk, social risk, privacy risk, subjective norm, intention to use social networking sites, and intention to recommend others to use social networking sites. The obtained Chi-square value for the measurement model was 694.29 with 332 DF and a p-value of .068. The normed chi-square value χ² / (df=332) was thus 2.38. The acceptable value of normed chi-square to show fit according to Schumacker and Lomax (2004:238) was less than 5. Other fit statistics showed good fit. The RMSEA was 0.04, TLI was 0.95, and GFI was 0.92, while NFI was 0.96.
For a model to be regarded as fit, the Tucker-Lewis Index (TLI) and the Normative Fit Index (NFI) need to be 0.95 or more than this figure, and the Root Mean Square Error of Approximation (RMSEA) has to be below 0.6, while the Goodness of Fit Index (GFI) needs to be 0.9 or more (Hu & Bentler, 1999:37;Baumgartner & Hombur, 1996:153). Fit outputs contain a large array of model fit, but this study reports only the commonly reported fit statistics.
Cronbach alpha was calculated in order to ascertain scale reliability. Constructs are considered reliable when the reliability coefficient is 0.7 or greater (Hu & Bentler, 1999:37). Results in Table 2 show that all constructs used in this study had reliability of above 0.70.
The measurement model's convergent validity was tested using factor loading, composite reliability (CR) and Average Variance Extracted (AVE). Convergent validity was demonstrated since all the items in Table 2 displayed a factor loading which is above 0.50 (Hair, Black, Babin & Anderson, 2010). CR results on Table 2 show that all the items displayed values which exceed 0.70, demonstrating convergent validity (Hair et al., 2010). The values of AVE for all the items as shown in Table 2 exceed 0.50, showing good convergent validity (Fornell & Larcker, 1981:45).
The model's discriminant validity was tested using, the Maximum Shared Squared Variance (MSV), the Average Squared Variance (ASV) and the square root of AVE values. According to Hair et al. (2010:86) and Fornell and Larcker (1981:46), the AVE values should be greater than MSV and ASV values. The square root of AVE should be greater than interconstruct correlations. MSV and ASV values in Table  2 show that all items are less than AVE values, and Table 3 also shows that the square roots (shown as the bolded diagonal figures) are greater than interconstruct correlations.  The square root of AVE values is represented diagonally (in bold)

Structural model
The results of PLS path coefficients are shown in Fig 2. The obtained results in Figure 2 show that all the paths are statistically significant. According to the results, 18% of the difference in perceived usefulness is revealed by the structural model, 40% of the difference in attitude is revealed by the structural model, 22% of the difference in intention to use is revealed by the structural model, and 18% in of the difference intention to recommend is explained by the structural model. This shows that the model offers a good explanation of the use of online sites for trip organising.
The PLS results in Figure 2 report that perceived risk (b=-.12, p<.05) negatively affects perceived usefulness, thereby reducing its perceived usefulness. Perceived risk (b=-.28, p<.01) negatively influences attitude towards the use of social networking sites for trip organisation. Hypotheses 3 and 4 are therefore supported. The results also show that perceived usefulness (b=37, p<.001), subjective norm (b=15, p<.01) and perceived behavioural control (b=28, p<.01) positively influence the attitude towards the use of social networking sites. Hypotheses 1, 2 and 5 are thus supported.  Perceived behavioural control (b=18, p<.01) was found to have a significant influence on intention to use social networking sites, implying that hypothesis 6 is supported. The results also indicate that attitude (b=42, p<.001) significantly influences both the intention to use social networking sites and to recommend others to use social networking sites for trip organisation (b=13, p<.05). The use of social networking sites thus support hypotheses 7 and 8. The results therefore provide evidence for the support of the proposed model. Part of the model was hierarchical, since some factors (perceived usefulness and perceived risk were made up of various dimensions (Wetzels, Odekerken-schroder & Van Oppen, 2009). The first two factors included in this study, perceived usefulness and perceived risk, have dimensions. Two dimensions exist for perceived usefulness (functional and social benefits and three dimensions for perceived risk (privacy, social and time risk). Thus, the importance of each dimension in building the second order constructs was also considered. From Figure 2 it is deduced that privacy risk (0.89) displays a value which is statistically different from the other dimensions (social and time risk), therefore privacy risk is more important in producing the second order construct perceived risk. Functional benefits (.90) also display a value which is statistically different from social benefits (.57) and is therefore more important in building the second order construct, perceived usefulness. Against this background, Table 4 provides a summary of the hypotheses testing.

DISCUSSION AND CONCLUSION
The use of social networking sites for trip organising is increasing daily (Lange-Faria & Elliot, 2012:197), primarily because it provides an ideal platform for users to interact and share their travel experiences by posting comments, sharing pictures and videos. The model proposed in this study highlights that perceived risk negatively impacts both perceived usefulness and the attitude towards the use of social networking sites for trip organisation. This conclusion concurs with Featherman and Pavlou's (2003) study which postulated that perceived risk adversely affects online sites' perceived usefulness. However, the influence of perceived risk on attitude was found to be greater than on perceived usefulness. This shows that if people perceive that there is a risk associated with using some social networking sites, they develop a negative attitude towards the sites. The model also illustrates that among the facets of risk (time, social and privacy risk) used for this study, privacy risk (.89) displayed a value which is significantly different from the other two facets, implying that it has the greatest impact on building the second order construct (perceived risk). Time and social risk displayed values which are almost close to each other in the building of perceived risk, thus it can be reported that these two types of risks have the same weight in the construction of perceived risk construct. This implies that people are concerned about the privacy risk associated with social networking sites, as compared to social and time risk. This conclusion is also similar to the findings of Featherman and Pavlou (2003) establishing that people are not as much concerned about social risk when using online sites for purchasing products.
The proposed model in Figure 1 also confirms the conclusion that perceived usefulness is one of the key factors which influence an individual's attitude towards the use of social networking sites for trip organisation. In addition, the two types of benefits used for this study (functional and social benefits) display values which are significantly different (.90 and .57 respectively). This implies that functional benefits are more important in building the second order construct (perceived usefulness) than social benefits. This means that individuals consider the functional benefits of using social networking sites for trip organising more than the social benefits. This finding concurs with the findings of a previous study conducted by Lopez et al. (2011), reporting that functional benefits have the greatest impact on building perceived usefulness.
The research also confirms that both perceived behavioural control and subjective norm influence one's attitude regarding the use of social networking sites. This implies that if people are of the opinion that they can successfully use a particular social networking site for trip organising, they will develop a positive attitude towards using the site. Similarly, advice and recommendations from close friends (subjective norm) can also positively influence the attitude of an individual towards the use of social networking sites. However, the effect of subjective norm on attitude is moderate, as this is shown by a probability value of .05. The possible reason might be that when using social networking sites, individuals are mostly on their own, hence the pressure from others on attitude is weaker. This finding is consistence with the findings of Akman (2014) and Hocevar et al. (2014) who also concluded that perceived behavioural control is important in influencing one's attitude towards social networking sites.
The research also reports that perceived behavioural control influences one's intention to use social networking sites with regard to travelling. When an individual perceives that he/she is capable of and certain that he/she can use social networking sites for trip organising, the intention to use the sites will increase. This supports the general rule of the TPB theory which states that when perceived behavioural control is greater, the person's intention to perform behaviour becomes stronger. Results from the model proposed for this study also confirm that the attitude towards social networking sites significantly influences the behavioural intention. This result is also in line with the rule of TPB which states that, the person's intention to perform behaviour becomes stronger when the attitude is more favourable. It therefore implies that if an individual has a positive attitude towards social networking sites, he/she will be interested in using them for trip organising and will not hesitate to recommend the site to others. This is consistent with the findings of Peslak , Cecucci and Sendall (2012), Zoonen et al. (2014) and Wei et al. (2015) who postulated that the attitude towards social networking sites is positively associated with the intention to use social networking sites.

RECOMMENDATIONS
The results of the study confirm that perceived usefulness exerts the greatest influence on the attitude towards the use of SNS for trip organisation. It is therefore recommended that tourism destination marketers or managers must upload valuable travel information such as the attractions found in the particular destination (through pictures or videos), how to get there (directions), as well as service offered at the destination on the social media accounts so that when travellers are seeking travel information, they can quickly access it to fulfil the functional benefit part. For example, tourism social media marketers can create a link from their Facebook or Twitter accounts to the company's website to inform travellers of services that the company can offer. Negative comments on the social networking site accounts of companies should also be managed in a fast and professional manner. The reluctance to do so has the potential to instil a negative attitude towards the use of social networking sites for trip organisation. A company can have a professional social media employee who is responsible for monitoring the account such, that when negative comments are posted, they can be quickly attended to. One way of managing negative comments is to communicate the issue to the leadership and respond on how the matter will be rectified, and remember to do it in a pleasant, positive and playful tone so that the situation is not aggravated.
Perceived risk has also a negative influence on attitude towards usage of social networking sites for trip organisation. Tourism social media marketers must design and develop systems that are transparent and to ensure that information found on social networking sites is trustworthy. Previous studies have shown that members' perceived risk increases when they perceive that their privacy can be abused through using online technologies (Chen, 2013). It is suggested that appropriate policy should be put in place to guard against privacy breach and to avoid further abuse. It is the duty of social networking site operators to ensure people's privacy online is protected in order to attract more users to the site as well as to maintain the users.
The results furthermore suggested that the attitude is another crucial factor which affects the intention to use social networking sites with regard to trip organisation. Hence, social networking sites service providers should come up with ways of improving their services in order to satisfy the demands of users so that they will continue using social networking sites for trip organisation. It was also found that perceived behavioural control influences the attitude towards the use of social networking sites, as well as the intention to use social networking sites for trip organisation. Therefore, social networking site service providers should ensure that all applications on social networking sites are user friendly. This can be achieved through making the social networking site mobile compatible, since more people are now using their mobile phones to access the Internet. If a company's social networking site cannot be accessed on mobile phones, then the company has to create a mobile version of that social networking site with the aid of web-based mobile website builders. Social networking site service providers should also ensure that there is effective navigation from one social networking site feature to another.

LIMITATIONS OF THE STUDY
The limitations of this study includes: Firstly, the population of the study comprised of individuals in only one province of South Africa and this might not give a true version of actual social networking sites usage in South Africa. However, the results of the study received some support from previous studies, thus generalisability might not be a problem. It is therefore recommended that future research should focus on a larger sample, if possible covering major towns in South Africa. Secondly, the study employed a cross-sectional approach where data was collected once and analysed. Future research may use a longitudinal approach to establish the actual use of social networking sites. Since the majority of people who participated in this survey were aged between 17 and 22 years, this might not give a clear picture of how older people view the use of social networking sites with regard to travelling. Future research might need to target the older age groups and a comparison can then be drawn.

CONCLUSION
The main aim of this study was to assess the influence of attitude and its precursors on the usage of social networking sites for trip organisation. It is concluded that perceived usefulness measured by functional benefits and social benefits is the key factor which influences attitude towards the use social networking sites for trip organisation. It is the responsibility of destination marketers to provide all the necessary or valuable information on their social networking site accounts, in order to encourage travellers to use social networking sites. On the other hand, perceived risk reduces social networking sites' perceived usefulness and negatively affects the attitude towards the use of social networking sites for trip organisation. Social networking site service providers should put stringent measures in place. Perceived behavioural control and attitude both influence the intention to use social networking sites. The article contributed to the existing literature by providing an extended TPB model which can be used to analyse individuals' attitude towards the use of social networking sites for trip organisation.