0 Contagious SWOM:
Effects of Tie Strength, Language Features and Valence of Reactions on Backpacker’s Risk Perception towards Malaria within SNS Communities
Gabriëlle Berry Dr. Gert-Jan de Bruijn
10362541 03-02-2017
Master’s Thesis Word count: 7489
Graduate School of Communication
1 Abstract
The present study examined a relatively understudied group, namely backpackers. Backpackers are prone to health risk taking behavior all for the authentic experience of independence during their travels. Therefore this study focussed on how the type of language used in a message and the valence of the reactions of other community members can moderate the effect of the tie strength between backpacker community members on the formulation of their SWOM engagement and risk perception regarding the use of malaria pills. An
experiment was carried out in the Dutch Facebook backpacker group called “Azië
backpackers”. Respondents consisted out of Dutch Facebook users who are members of the “Azië backpackers” community. The results were analyzed by using Univariate Anova of Variance. The findings of this study suggest that language features, such as controlling or unforceful language, can successfully increase the risk perception of backpackers regarding malaria pills. Neither tie strength between community members nor the valence of reactions had a significant effect on risk perception nor SWOM engagement. Implications of the findings suggests that health organizations have the option to insert themselves into conversations within backpacker communities. Organizations can then choose between unforceful or controlling language to formulate a message to increase risk perception. But then there is a risk that backpackers will be more apprehensive to accept information due to the persuasive nature of messages.
Keywords: SWOM engagement, SNS, risk perception, tie strength, backpacking, valence of reactions, type of language
2 Introduction
The emergence of social media has allowed for the formation of online backpacker
communities as a playground for backpackers to spread and gather information to others like them. All for the authentic experience of foreign faraway cultures and its native people on their journey (Cohen, 2003). Backpacking is seen by many backpackers as a rite of passage from teenage years to adulthood. By travelling to unfamiliar destinations abroad backpackers must prove to others and most importantly to themselves that they can be independent on their journey. That backpacking is an opportunity for a life transition without guidance from others (Turner & Turner, 1978).
But even the best of backpackers needs help or guidance sometimes. In the age of social media backpackers can look towards other backpackers for their help unrestrained by either time or space. In this manner backpackers can maintain their feelings of independence which is seen as the main goal of backpacking by many backpackers in the first place. As a result backpackers value the opinions and experiences of other backpackers on social network sites (SNS) over the professional advices from health practitioners about health risks at certain destinations. This poses a problem for health organizations. In general backpackers, mainly consisting out of young adults and millennials, have a lack of knowledge about potential health risks while travelling. Also because the advices of other backpackers are not always realistic or safe for someone’s health. Thus subjecting backpackers to a higher risk of
potentially getting sick or injured (Dahlman & Stafström, 2013). This begs the question how can health organizations positively steer the conversations between backpackers on SNS to increase their risk perception about health risks? Especially when the presence of health organizations is unwelcome on SNS and its communities (Fournier & Avery, 2011).
Previous academic research on SNS claims that the effect of a social word-of-mouth (SWOM) message depends on the intensity of the considered importance of a social tie
3 originating from someone’s social network (Steffes & Burgee, 2009). The stronger the tie to another person, the more reliable and credible a message is perceived to be (Brown & Reingen, 1987). But in order for a SWOM message to be successful it first has to be seen as not being a threat due to being too persuasive in nature. To reduce the threat level of a message language features can be altered to change the type of language that is being used in a message (Jenkins & Dragojevic, 2011). The believability of SWOM information also
depends on the valence of the information. Negative valence is more influential in influencing the credibility and believability of SWOM information than positive valence (Koo, 2015). Resulting in a greater impact on the risk perception of backpackers about health issues, especially when the information originates from a strong tie (Lee, Rodgers & Kim, 2009).
Backpacking is a relatively understudied field within Communication Science. While the body of academic research on SNS communities is growing, the literature on backpacking communities remains absent. Thus allowing for a research gap to occur. Which seems odd, because there is an academic consensus that social media functions as one of the primary sources of information that people first come into contact with (Giordano & Giordano, 2011; Kaplan & Haenlein, 2010; Brossard & Scheufele, 2013). Social media reshaped the way in which people look for information. Mainly because of its popularity and easy accessibility teens, young adults and millennials often first look on social media, such as Facebook and Twitter, for health information before they look elsewhere (Vance, Howe & Dellavalle, 2009). This research aims to resolve this gap. By offering a societal solution for health practitioners as to how they can increase backpackers’ risk perception without coming across as intrusive or unwelcome on SNS. The solution lies in steering the social word-of-mouth (SWOM) conversations between community members to the advantage of health
4 while simultaneously behaving in a responsible way to reduce sickness or injuries
forthcoming out of irresponsible behavior.
This research also wants to add an academic understanding about how SWOM
conversations on SNS can influence users their risk perceptions and engagement with a topic. To better understand how social network ties can increase or decrease backpackers their risk perception about health related issues, which in turn can have consequences on the decision-making and offline daily lives of backpackers. It isn’t enough anymore to just simply
understand how conversations play out between users on social media. Instead research also must learn how to influence these conversations expressed in an SNS environment to increase risk perception.
This thesis aims to add an understanding about the dispersion of SWOM on SNS and how this can influence the risk perception and SWOM engagement of backpackers towards health related issues. Therefore in this research I will focus on type of language and valence of reactions to show the moderation on the tie strength between backpacker community members on SNS on the formulation of backpackers their SWOM engagement and risk perception towards malaria pills. First I will begin by looking at SWOM and tie strength, followed by the role of type of language as a message characteristic as well as the valence of SWOM reactions on the formation of health risk perception and SWOM engagement.
Theoretical Framework
Word-of-mouth. Word-of-mouth (WOM) communication has been claimed to be more influential than other types of communication, such as media, because of the role of an
interactive two-way dialogue instead of an one-way flow of information (Katz & Lazarsfeld, 1955). Also defined as informal and oral face-to-face communication between people, WOM is regarded to be of a non-commercial nature regarding services, products, brands or
5 Social word-of-mouth. Unlike WOM, with social network sites word-of-mouth
(SWOM) consumers can communicate unrestrained by time and space allowing them to communicate in a manner that is most beneficial to them (Sun, Youn, Wu & Kuntaraporn, 2006). SWOM is defined as written communication on social networking sites (SNS)
broadcasted from one person to many receivers in a non-face-to-face interaction to someone’s online social network. A definition that has reached to most consensus in academic research is that users create profiles to make connections to other SNS users and to interact with SNS content and shared information (Eisingerich, Chun, Liu, Jia & Bell, 2015).
Social environment. Studies on traditional WOM researched characteristics of the social networks of providers and receivers of WOM information. Within someone’s social network there is a wide range of relationships and social ties. These social ties range from non-existent ties (strangers without shared interests) to weak ties (acquaintances) and strong ties (friends and family). The strength of a social tie is determined by the intensity of the considered importance of that relationship to a person (Steffes & Burgee, 2009). Consumers often rely on strong tie connections for the provision of information. Because information provided by strong tie connections is perceived to be reliable, thus being more persuasive (Brown & Reingen, 1987). If the provided information is based on the personal experiences and opinions from strong tie connections then the information is also seen as being more credible (Gilbert & Karahalios, 2009). Stronger ties also initiate more awareness than weaker ties and are more appreciated by others (Chu & Kim, 2011). For backpackers this would mean that strong tie connections can initiate more awareness in evaluating SWOM information regarding potential health risks, resulting in a higher risk perception.
However, research has shown that weaker ties also play an important role within someone’s social network. Because weak ties are important for the provision of
6 allow for information to flow between social networks on different SNS (Koo, 2015). Also weak ties are more powerful in influencing future preferences, while strong ties are more influential in providing future recommendations (Zhao & Xie, 2011). Therefore based on previous academic research it is expected that backpackers will be more persuaded by strong tie connections than by weak tie connections.
H1: Backpackers will develop a higher risk perception towards malaria from SWOM information coming from SNS users to whom they have strong tie connections than from SNS users to whom they have weak tie connections.
Prior research suggests that the impact of SWOM of social ties on SNS users their decision making depends on its main determinants. The effectiveness of SWOM depends on determinants such as the impact of source credibility and message relevance. For the source of a message to be seen as credible, the sender of the information must be deemed trustworthy and have the required expertise to make judgements about certain products, information, services, organizations or brands (Huang & Chen, 2006). When the receiver of the
information deems the sender trustworthy, meaning that the sender isn’t biased or is having an alternative motive for helping the receiver, and having enough expertise then the receiver will consider the credibility of the information to likely be good (O’Reilly, MacMillan, Mumuni & Lancedorfer, 2016).
Furthermore, SNS users assess the message relevance of SWOM based on how similar the source is to them in terms of personality and interests. The first thing that a receiver does when coming into contact with new information is judge the relevance of that information for them or their goals. Therefore any communication or information must be deemed relevant to a receiver before it can be effective. Assessment of the relevance of a message based on
7 personality similarity refers the contemplation of a receiver about how similar the sender of the information is to them based on their personality, opinions, social connections, interests and background (O’Reilly et al, 2016). In essence this means that an attractive, similar, stronger tie or most likeable sender or source of information is most likely to have the largest impact on a receiver when compared to a lesser attractive source (Turner, 1991). This occurs because people tend to like other people better when they perceive others to be similar to themselves in regards to opinions, interests and personality. In turn they perceive the information provided by these similar senders to be more relevant to them (Thompson & Malaviya, 2013).
While SNS users judge the relevance of messages based on shared interests and personalities, backpackers also judge the message with a key motivation in mind.
Backpackers use their connections to fellow backpackers in backpack communities on SNS for the sharing of information and gaining knowledge that they can use for their journey (Zhang, 2014). Gathering information is a key motivation for backpackers to interact with each other. Through SWOM interaction concerning the gaining of knowledge backpackers can communicate with each other when they are away from home. Therefore SWOM has two main functions. Firstly, SWOM interaction is not confounded by either time or space. During their travels backpackers don’t adhere to a vast schedule. Therefore the flexibility of SWOM communication allows for the gaining of knowledge through SNS interaction in a manner that is most suited to them (Zhang, 2014). Secondly, backpackers gain knowledge through SWOM interactions with fellow backpackers (Murphy, 2001). Backpackers form connections to each other to “recreate” the presence of home through the sharing of experiences and opinions about their preparations or their backpack journey (Adkins & Grant, 2007). In this manner backpackers provide each other with information in a SNS community of strangers. Thus they
8 don’t have to rely on outsiders, meaning people who aren’t members of the backpacker
community.
But contemporary research focused on tie strength still adheres to the traditional definition of the concept and its categories. Thus allowing a research gap to surface because in the age of SNS SWOM categories can differ from WOM categories of tie strength. With the rise of social media a change occurred in the displays of meaningful interactions and formations of relationships. Tie strength can increase over time through SNS communication through either directed communication (one-on-one direct communication) or passive
communication (reading about connections on SNS without directly contacting them), (Burke & Kraut, 2014). But tie strength on SNS can be hard to measure. Due to the fact that
traditional communication is operationalized differently on SNS (Gilbert & Karahalios, 2009). Because how is tie strength between community members expressed on SNS? Is it shown through the sharing of messages between users, the quantity of likes and comments on a post, the exchange of information or the simultaneous appearance with friends in a
photograph? Engagement expressions to other users or community members is often
measured through participation. Where users use their SNS profile to engage with their social network connections through interpersonal communication (Ashley & Tuten, 2015). Users can passively consume SNS content and information or they can actively engage by contributing user-generated content (Hutton & Fosdick, 2011). But SNS engagement is unrestrained by either time or space. This means that participation or interpersonal
communication doesn’t synchronously have to occur on SNS for two people to interact for their relationship to grow closer. In this manner SNS is used to nurture online relationships and strengthen social ties (Burke & Kraut, 2014).
But the nurturing of online relationships within backpack communities on SNS serves as a means to an end. Through the exchange of information community members both obtain
9 and exchange information to help each other to reach a goal (Wang, Yeh, Chen & Tsydypov, 2016). The stronger the tie to other community members the stronger the community feels the need to help its members to reach a common goal (Coleman, 1990). Through social
engagement within an online community information can be shared to reach a common goal, namely gathering information (Murphy, 2001). In the age of SNS this would mean that through the posting, liking, commenting and sharing of information community members socially engage with other members so that together they can reach a goal. This could mean that an increase in tie strength can occur over time through engaging with other community members over shared interests through the exchange of information. Resulting in stronger feelings of belonging and ties to both the online community and to other members.
H2: Backpackers will feel a higher need to engage in SWOM from a strong tie connection than from a weak tie connection.
Most importantly based on prior research it is expected that through the guidance of fellow community members, backpackers can collect and share information as a community instead of independently in the age of SNS. Therefore backpackers don’t need guidance from “outsiders” outside the backpack community on SNS. In this manner backpackers can
maintain their feelings of independence and freedom. Which are seen as the main goal for backpacking in the first place (Turner & Turner, 1978). By increasing their engagement towards achieving common goals by helping fellow community members, backpackers can receive social benefits out of their SWOM participation forthcoming out of their membership of an online backpack community (McWilliam, 2000). Because due to these social benefits fellow community members will be more inclined to help you achieve your goal if you help them beforehand or in return.
10 Type of language. In order for social ties to be influential in SWOM conversations within online backpack communities, the formulation of the message that is being spread also has to be kept in mind. Because to understand the inner workings of online backpacker communities, academic research has to investigate who is participating in SWOM
conversations and their motives, but also as to how SNS users are communicating through the usage of message characteristics in the first place. In the age of SNS messages or
recommendations are masqueraded in SWOM comments or posts to reduce a user’s resistance against persuasion. In this manner the message won’t be perceived as a persuasion attempt, thus lowering resistance against the message (Reijmersdal van et al, 2016; Pfeffer, Zorbach & Carley, 2013). According to the politeness theory not only will these messages be perceived as direct threats to someone competence but also towards their freedom to choose (Brown & Levinson, 1987). As a defensive mechanism to protect or preserve competence or freedom, SNS users can decide to ignore the message or comments all together (Jenkins & Dragojevic, 2011). In other words, this would mean that if SWOM is perceived to be unjustified or as unruly then users are not persuaded by the content of these messages even if the message comes from a (strong) social tie. Thus increasing their reactance towards a SWOM message.
According to the theory of psychological reactance people in general have a basic need for autonomy. If their behavior is threatened by an outside force then people experience reactance as a result (Brehm, 1966). Reactance is an emotional state best described as a negative and aversive attitude towards messages that people want to avoid. Senders of information can attempt to reduce the threat by including message characteristics to reduce reactance to the persuasive nature of messages (Burgoon, Alvaro, Grandpre & Voulodakis, 2002). Language features are the most common used message characteristic to reduce the persuasive presentation of messages (Dillard & Shen, 2005). When confronted with a persuasion attempt people often have one immediate reaction, namely reactance. It is an
11 attempt of the receiver of the message to resist persuasion and the change that a message might bring forth (Knowles & Linn, 2004). The persuasive nature of a message can be masked by language features to induce resistance against persuasion. Persuasive arguments given in a message are then perceived to have a lesser forceful impact on someone’s autonomy. For instance, forceful language can be employed to increase or decrease the
persuasive nature of arguments given in a message (Dillard & Shen, 2005). Forceful language consists out of two types of basic languages, namely controlling and demeaning language. By employing controlling language a message orders its readers to listen to commands as to how people should behave and think (Miller, Lane, Deatrick, Young & Potts, 2007). However with demeaning language a message criticizes the competence of its readers. By questioning whether or not people can make responsible decisions or to perform the advocated behavior, demeaning language leads to more resistance towards messages. Even if recipients agree with the content of a message, they are unjustly being criticized which threatens their competence (Jenkins & Dragojevic, 2011). Due to the persuasive nature of both controlling and
demeaning language, people feel threatened and as a result resist the content of a message completely. Simply by the fact that people don’t like to be told what to do because it threatens their freedom (negative face) and because it disrespects people and questions their
competence to be able to make good decisions (positive face), (Brown & Levinson, 1987). Although SWOM messages can generally be perceived as persuasive, unforceful language characteristics can attempt to reduce the threat level of the formulation of these messages to one’s freedom or competence. By using unforceful language messages emphasize that people are competent to perform certain behavior or to hold certain opinions. Moreover unforceful language says that people are free and have the right to choose (Jenkins &
Dragojevic, 2011). Because unforceful language doesn’t threaten people’s need for autonomy and independence, it induces reactance towards messages. Because unforceful language
12 masks the persuasive nature of these messages. Information formulated in these messages can then persuade backpackers to act responsibly and to gain knowledge regarding health risks. Thus increasing their risk perception. Ergo, unforceful language would be the most successful language type in reducing the persuasive nature of messages. Consequently lowering
resistance towards messages coming from social ties originating from someone’s SNS network. Therefore it is expected that backpackers will be more persuaded by unforceful language than by controlling or demeaning language because it triggers less resistance.
H3: Backpackers will develop a higher risk perception from messages employing unforceful language than from messages that employ controlling or demeaning language.
Valence of reactions. Within backpack communities on SNS information is shared and evaluated through SWOM by its members. With the emergence of social media users can connect with other users and share information, opinions and experiences with their social connections all over the world (Evans, 2010). Therefore the role of the SNS user also includes the shaping of the experiences of other users with the sharing of SWOM. Due to the easy accessibility and large reach of SWOM, SNS is often seen as the first source of information that users come into contact with when they want to find information about locations, services, products, organizations or brands (Fulgoni & Lipsman, 2015). This makes SWOM on social media nowadays potentially the most influential information source on the
formation of opinions, intentions and perceptions of other SNS users (Evans, 2010). Thus in the current day and age SNS users have the most influence on the formation of perceptions of other SNS users.
13 But the believability of information provided via SWOM depends on the knowledge that SNS users might have about the subject at hand. Users who have no knowledge are most easy to influence in comparison to users who do have the knowledge. Negative SWOM is hereby the most influential (Bambauer-Sachse & Mangold, 2013). But SWOM can be deceitful. Not all SWOM comments or reviews are truthful because SNS users can post comments without having any experience or knowledge about the subject at hand. Often these comments are negative in nature (Anderson & Simester, 2014).
Previous studies show that negative SWOM holds a higher value than positive SWOM when it comes to the credence of experiential services or goods. A reason for negative valence to be more influential in SWOM is that with buying or trying experience goods the SNS user has more to lose. Due to the fact that judgements can only be made about the offered good after someone experiences it. Because everyone judges the good or service differently and because experience goods have no fixed attributes that someone can evaluate before using it (Koo, 2015). The effect of negative valence on information, events, experiences and emotions is more influential than positive valence. This is also known as the negativity bias (Rozin & Royzman, 2001). Negative valence expressed in SWOM comments by other community members will therefore have a greater impact on influencing backpackers their judgements, attitudes and opinions (Lee, Rodgers & Kim, 2009). Especially when the intention to perform a certain behavior is high (East, Hammond & Lomax, 2008).
When confronted with negative SWOM comments or reviews SNS users look for more information from other sources. As users have the opportunity on SNS to closely observe social connections within social networks or SNS communities they regard the provided information from those strong tie connections to be more truthful and persuasive, especially when the information is negative in nature (Zhao & Xie, 2011). Strong ties are therefore most influential in providing information. But when confronted with positive
14 recommendations or comments online people regard the information as normal (Chiou & Cheng, 2003). Because only satisfied people post positive comments and have more favorable attitudes (East, Hammond & Lomax, 2008). Building on the findings of previous research it is expected that negative valance about provided information via SWOM will carry more weight and be more influential in negatively influencing backpackers their risk perception than positive valence.
H4: Backpackers will develop a lower risk perception from SWOM holding negative valence on SNS than from SWOM holding positive valence.
Risk perception. Previous studies have shown that backpackers are subjected to a higher risk of getting injured or sick while traveling through a foreign country. All due to the fact that backpackers relatively travel on a low budget and individually, they are prone to risk taking behavior all for the experience of authenticity and freedom (Murphy, 2001). Another factor is that backpackers often have a lack of knowledge about potential health risks present in a country (Dahlman & Stafström, 2013). To gain knowledge backpackers rely on
experiences of others to maintain a feeling of authenticity.
Risk taking behavior is prone to the perceived experience of the sender of the
information regarding backpacking and perceived (health) risks. Backpackers rely extensively on the experiences of others similar to them to make responsible decisions. If they perceive that employees from an institutional organization or support, like GGD, don’t have the relative experience or aren’t similar to them then backpackers disregard the given advices because they find them to be unreliable (Hunter-Jones, Jeffs & Smith, 2008). Rejection of institutional support given by licensed health employees alongside inexperienced and often young backpackers makes it that this group is prone to risk taking behavior. Instead
15 backpackers look towards connections from their social network to provide them with reliable information. This phenomenon is also known as the contagion theory. According to this theory people accept and take over the opinions, behaviors and attitudes from people that they have strong tie connections with (Scherer & Cho, 2003). Therefore the stronger the tie, the stronger the effect of another person’s beliefs on backpacker’s perception of (health) risks.
To positively influence the perception of health related risks key opinion leaders within SNS communities should be engaged by health organizations to positively endorse responsible behavior. But organizations should not overtly express their presence on SNS because it will be seen as intrusive and unwelcome. SNS and its communities are perceived to be for people and not for organizations or brands (Fournier & Avery, 2011). Health
organizations shouldn’t rush to invade SNS. Instead organizations should let the conversation play out between people by endorsing key opinion leaders. Thus increasing precautious behavior regarding health risks while backpacking. Engaging with key opinion leaders and people that have strong ties to other people in a social network as endorsers of positive and responsible health behavior should increase precautious behavior in their social network (Scherer & Cho, 2003). Because key opinions leaders are strongly tied to their social network they have the largest social capital within SNS communities and are therefore seen as reliable by other members. Thus increasing the perceived risk that people might associate with certain behaviors (Kelly et al, 1991).
Method
To investigate the research question an experimental approach has been chosen to investigate how tie strength to social connections can affect backpacker’s risk perception and SWOM engagement with a message and to investigate the moderation of valence of reactions and different types of language characteristics used in the formulation of messages.
16 Specifically, this experiment tested the effects of strong and weak ties on risk perception and SWOM engagement, as well as the moderation via type of language and valence of reactions.
Research design and variables. This study was designed as a 2 (tie strength: weak, strong) x 3 (type of language: unforceful, demeaning, controlling) x 2 (valence of reactions: positive, negative) between subjects design. This design focused only on the main effects of the 12 conditions and not on the interaction effects. A 7-point Likert scale was used to measure the perceptions of respondents to different senders of information while using different types of languages as formulated in the post. The 7-point Likert scale ranged from “1= helemaal mee eens” to “7= helemaal niet mee eens” as formulated in Dutch. Translated to English this would mean that the 7-point Likert scale ranged from “1= completely agree” to “7= completely do not agree”. At the beginning participants had to answer general questions in the online experiment. Questions such as what is your gender and what is your age. Next, participants were shown a fake Facebook post containing information about malaria. The Facebook post was manipulated so that participants were shown one of the twelve conditions of the variables tie strength (strong or weak), valence of reactions (positive or negative) and type of language (controlling, unforceful or demeaning). Through randomization participants were assigned to one of the twelve conditions as mentioned above. Depending on the assigned condition participants saw that the sender of the information (weak tie: GGD op reis or strong tie: Michael Kroon) wrote a personal introduction and general information about malaria followed by advice for safe backpacking in Asia through malaria risk areas (formulated by using either controlling, demeaning or unforceful language) and ending with reactions of other fictional members of the Facebook group “Azië backpackers” on the post in the comments section (formulated either positive or negative).
Preceding the randomized assigning to one of the twelve conditions, participants were shown a brief message containing information about the following block and asking
17 participants to read carefully because questions would follow. Per condition five blocks were presented to the participants. Per block questions were formulated to measure each variable. Questions in the blocks of the variables “tie strength” and “type of language” were presented in a random order. Additionally questions in the blocks of the variables “valence of
reactions”, “SWOM engagement” and “risk perception” were presented next.
Selection research unit. Facebook was chosen as the main recruitment social media platform because Facebook is popular among backpackers as a source of information and as a platform to connect with other backpackers. The public Facebook group called “Azië
backpackers” was chosen as an appropriate online community to recruit respondents. The main reason for choosing “Azië backpackers” is that this particular Facebook group is the largest public backpack group in the Netherlands that is focused specifically on Asia. Travelers and specifically backpackers are confronted with health risks when travelling to Asia. Even if the traveler himself is unaware. For backpackers health risks such as malaria are often the cause of sickness or early departure back to the Netherlands. The Facebook group “Azië backpackers” is a Dutch public group founded by backpackers originating from the Netherlands. Both men and women are welcome to join the Facebook group. The group consists out of Dutch Facebook users who are interested in travelling to Asia in the nearby future, who are now travelling in Asia, who’ve been to Asia in the past, family members of travelers that are now in Asia or users that are just interested in the travel stories of
backpackers in general.
Procedures. Respondents were recruited via a post on Facebook in the public group called “Azië backpackers”. The post displayed a message asking potential respondents to partake in this experiment so that insights into health risk perceptions could be derived from the Dutch Facebook backpack community interested in Asia. The message also contained an incentive if respondents were to completely fill in the questions in the survey because then
18 they could win an e-book gift certificate. After reading the message respondents were asked to click on a link, directing them to the online experiment. The experiment lasted 50 days after its activation before it was deactivated.
Characteristics of research unit. Participation in this experiment was limited to only members of the “Azië backpackers” Facebook group whom are living in the Netherlands. The language of the survey was thus in Dutch. Only Facebook users could partake in the
experiment. In total 23.734 members of the Facebook group “Azië backpackers” could’ve seen the post. Out of this number a total of 652 members (response rate 2.75%) clicked on the provided link directing them to the online experiment (see tables 5, 6, 7 & 8 in appendix for descriptives). Answers from respondents were excluded from the analysis if respondents didn’t finish the whole survey or if respondents took longer than 120 seconds to answer a question about information given by health organizations about precautious behavior regarding malaria. As to ensure that respondents didn’t look up information.
Variables. The variable “tie strength” was operationalized in this experiment through the sender of the information contained in the fake Facebook post. GGD op reis, an official GGD account on Facebook, was used to operationalize weak tie strength. A fake profile “Michael Kroon” was created, a member of the “Azië backpackers” Facebook community, to operationalize strong tie strength. In the beginning of the post the sender had a brief personal introduction. Experiences were expressed with backpacking or backpackers and positive opinions regarding precautious behavior regarding malaria. Furthermore the variable “type of language” was operationalized via language characteristics used in the Facebook post when the sender formulated an advice over precautious behavior. The reader was confronted with information formulated in either unforceful, controlling or demeaning language. Saying that backpackers should contact GGD and swallow malaria pills when travelling in malaria risk areas. Also the variable “valence of reactions” was operationalized via fake reactions on the
19 Facebook post from fictional “Azië backpackers” community members. Formulated in either a positive or negative manner the reactions were about the opinions of the agreement level about the information or past experiences regarding malaria.
For the correct measurement of each variable several items were used in the analysis. The variable “tie strength” consisted out of 7 items. Whereas the variable “risk perception” consisted out of 33 items. Furthermore the variable “SWOM engagement” consisted out of 8 items, the variable “valence of reactions” consisted out of 6 items and the variable “type of language” consisted out of 18 items.
Results
In order to test the four hypotheses an 2x3x2 Univariate Anova of Variance was conducted. To test H1, backpackers will develop a higher risk perception towards health risks from SWOM information coming from SNS users to whom they have strong tie connections than from SNS users to whom they have weak tie connections, an Univariate Anova was carried out. The Univariate Anova had as its dependent variable “risk perception” (0 =
middle, 1 = high, 2 =low), its independent variable “tie strength” (0 = normal, 1 = strong tie, 2 = weak tie) and as its moderators “type of language” (0 = demeaning language, 1 = unforceful language, 2 = controlling language) as well as “valence of reactions” (0 = neutral, 1 =
positive, 2 = negative). The analysis showed a non-significant effect of “tie strength” on “risk perception”, F(1, 16) = .00, p = n.s. As can be seen in the table 1 in the appendix, the results from the analysis suggest that the tie strength between SNS backpacker community members does not lead to a higher risk perception towards malaria health risks. To conclude H1 has to be rejected because there is a non-significant effect of “tie strength” on “risk perception”.
Additionally to test H2, backpackers will have a higher need to engage in SWOM from a strong tie connection than towards a weak tie connection, an Univariate Anova was carried out. The Univariate Anova had as its dependent variable “SWOM engagement” (0 =
20 middle, 1 = high, 2 =low), its independent variable “tie strength” (0 = normal, 1 = strong tie, 2 = weak tie) and as its moderators “type of language” (0 = demeaning language, 1 = unforceful language, 2 = controlling language) as well as “valence of reactions” (0 = neutral, 1 =
positive, 2 = negative). The analysis showed a non-significant effect of “tie strength” on “SWOM engagement”, F(2, 13) = .00, p = n.s. As can be seen in the table 2 in the appendix, the results from the analysis suggest that the tie strength between SNS backpacker community members does not lead to a higher SWOM engagement on SNS. To conclude H2 has to be rejected because there is a non-significant effect of “tie strength” on “SWOM engagement”.
Furthermore to test H3, backpackers will develop a higher risk perception from messages employing unforceful language than from messages that employ controlling or demeaning language, an Univariate Anova was carried out. The Univariate Anova had as its dependent variable “risk perception” (0 = middle, 1 = high, 2 =low), its independent variable “tie strength” (0 = normal, 1 = strong tie, 2 = weak tie) and as its moderators “type of
language” (0 = demeaning language, 1 = unforceful language, 2 = controlling language) as well as “valence of reactions” (0 = neutral, 1 = positive, 2 = negative). The analysis showed a significant effect of “type of language” on “risk perception”, F(2, 16) = 3.42, p = <.001. As can be seen in the table 3 in the appendix, the results from the analysis suggest that both unforceful and controlling lead to a higher risk perception towards malaria health risks than demeaning language. This means that H3 can partly be accepted, because unforceful language significantly affect the risk perception of SNS backpacker community members towards malaria health risks higher as does controlling language.
Moreover to test H4, backpackers will develop a lower risk perception from SWOM holding negative valence than from SWOM holding positive valence, an Univariate Anova was conducted. The Univariate Anova had as its dependent variable “risk perception” (0 = middle, 1 = high, 2 =low), its independent variable “tie strength” (0 = normal, 1 = strong tie, 2
21 = weak tie) and as its moderators “type of language” (0 = demeaning language, 1 = unforceful language, 2 = controlling language) as well as “valence of reactions” (0 = neutral, 1 =
positive, 2 = negative). The analysis showed a non-significant effect of “valence of reactions” on “risk perception”, F(2, 16) = .00, p = n.s. As can be seen in the table 4 in the appendix, the results from the analysis suggest that the valence of reactions of fellow SNS backpacker community members does not lead to a higher risk perception towards malaria health risks. To conclude H4 has to be rejected because there is a non-significant effect of “valence of reactions” on “risk perception”.
Conclusion & Discussion
The present study empirically tested the extent to which “type of language” used in a message and the “valence of reactions” of other community members on a message
moderated the influence of “tie strength” between backpacker community members on SNS on the formulation of their SWOM engagement and risk perception towards malaria pills. This study found evidence that the language features employed in a message can increase the risk perception of backpackers towards malaria. But in contrast to what was expected both unforceful and controlling language features used in the formulation of a message increased risk perception. Therefore H3 can partly be confirmed. This can be explained due to the fact that controlling language raised resistance. By doing so it enabled backpackers to focus on the content of a message to motivate themselves to resist the message or by arming themselves with arguments to resist the message of the sender further (Knowles & Linn, 2004). But because the information wasn’t threatening their health if they failed to perform the desired behavior but only informing them, respondents were lesser motivated to disregard the
information. Instead respondents reappraised and scrutinized the information, thus still raising their risk perception (Riet van ‘t & Ruiter, 2013).
22 Moreover, this study found that the tie strength between social connections doesn’t significantly cause a higher risk perception or an increase in willingness to engage in SWOM on SNS. Thus refuting the theorization for H1 and H2, stronger ties don’t enable a higher risk perception or a higher need to engage in SWOM in comparison to weaker ties. Furthermore this study found that the valence of SWOM reactions of SNS community members on a message doesn’t significantly influence a backpacker’s risk perception towards malaria pills, thus refuting H4. Contradicting the findings of Evans (2010) that SWOM reactions of fellow community members on SNS are the most influential information sources on the formation of the risk perception of other community members.
Implications. The findings of this study have several implications for academic theory. Firstly, this study added to the non-existent academic literature on backpacker communities on SNS. Thus being the first to test which factors play a significant role within Dutch SNS backpacker communities on the formation of their risk perception towards malaria.
Secondly, this study also added to the literature on reactance and ways to counter or not raise reactance. The findings confirmed the findings of previous academic research on language features, namely that unforceful language features are best to not raise reactance to a message in the first place (Knowles & Linn, 2004; Dillard & Shen, 2005; Jenkins &
Dragojevic, 2011). The findings show that by not raising reactance respondents had lesser resistance to the offered information on malaria. However in comparison to the findings of Knowles and Linn (2004) raising resistance by the use of controlling language features can actually increase acceptance of information in a message to increase risk perception. Thus proving that raising resistance towards a message isn’t always a bad thing. It’s true that people don’t like to be told what to do. But this attitude evokes scrutiny by which people carefully think to process the information before refuting it.
23 Thirdly, a contribution of this study to theory is that this study contradicted the
findings of previous academic research in regards to tie strength. Traditional studies of tie strength proclaim that strong social ties cause for an increase in communication and appreciation of information in comparison to weak social ties (Steffes & Burgee, 2009; Brown & Reingen, 1987; Chu & Kim, 2011). The findings demonstrated a key difference between traditional measures of tie strength and SNS measures of tie strength by being non-significant. Thus proving that traditional measurement of tie strength is no longer accurate in the age of SNS. Because outing of tie strength between users differ on SNS. These outings are harder to measure because the very definition of tie strength differs on SNS. Therefore the traditional measuring of tie strength will no longer uphold on SNS. Meaning that new ways to measure the tie strength between SNS users have to be further researched in the future. By being non-significant the findings of this study also contradicted the findings of previous studies that the presence of organizations on SNS would be unwelcome in SNS groups that were built by users and meant for users (Fournier & Avery, 2011; Noort van & Willemsen, 2011). Thus opening up the academic floor for future discussions about factors that might come into play to make the presence of an organization (un)welcome on SNS on user-generated platforms or groups.
The findings have several societal and practical implications for backpackers and health organizations. Suggesting a practical application that health organizations can be present in SNS communities to spread messages and to interject themselves into the
conversations between users to give accurate and responsible information regarding malaria or other health risks without being seen by users as intrusive. This means that health
organizations can move away from their own official organizational or brand-generated platforms to user-generated platforms or groups to reach potential consumers or clients. Thus moving the conversation to them. Another practical implication of the findings regards the use
24 of controlling or unforceful language features by health organizations in the formulation of a message to increase risk perception. Information regarding malaria or other health risks can effectively increase the development of a higher risk perception whether it is formulated using either unforceful or controlling language features. This means that health organizations can have the option to formulate a message by telling people how to act accordingly because then people will feel threatened and will experience reactance. Because although people
experience reactance they will still develop a higher risk perception regarding health risks. More importantly, the findings also have societal implications for the SNS backpacker communities. Due to a higher risk perception backpackers might become prone to less health risk taking behavior, because they have a higher knowledge regarding health risks and
precarious behavior. In regards to malaria this would mean that backpackers will have a lesser chance of becoming infected. Also like every SNS community, backpacker communities on SNS have their own culture, norms and values (Johnson, 2001). The Dutch SNS backpack community is still untainted by the presence of “outsiders”. Only backpackers and SNS users who are interested in backpacking are now present in that community. If health organizations were to make themselves openly present within SNS backpacker communities this could potentially affect the effectiveness of information to increase risks perception regarding health risks in the future. This could mean that in time health organizations will find it harder to reach their target group because backpackers will be more apprehensive to accept information due to the persuasive nature of messages.
Limitations. After careful consideration it was decided that the experiment was to take place in the “Azië backpackers” Facebook group which consisted out of 23.734 members. The standard response rate normally is 20%, meaning that it was expected that around 4746
members were to partake in the experiment. This would have been enough because the
25 what was expected not enough respondents were willing to partake or complete the
experiment. Even though a lot of effort has been put in to gather enough respondents. This might have influenced the significance of the results due to the limited power behind the data-set. Therefore the findings of this research might be biased and are in need of more research in the future.
Reference List
Adkins, B. A., & Grant, E. L. (2007). Backpackers as a community of strangers: the
interaction order of an online backpacker notice board. Qualitative Sociology Review, 3(2), 188-201.
Anderson, E. T., & Simester, D. I. (2014). Reviews without a purchase: low rating, loyal customers and deception. Journal of Marketing Research, 51(3), 249-269.
Ashley, C., & Tuten, T. (2015). Creative strategies in social media marketing: an exploratory study of branded social content and consumer engagement. Psychology and
Marketing, 32(1), 15-27.
Bambauer-Sachse, S., & Mangold, S. (2013). Do consumers still believe what is said in online product reviews? A persuasive knowledge approach. Journal of Retailing and
Consumer Services, 20(4), 373-381.
Brehm, J. W. (1966). A theory of psychological reactance. New York, NY: Academic. Brossard, D., & Scheufele, D. A. (2013). Science, new media and the public. Science,
339(6115), 40-41.
Brown, P., & Levinson, S. C. (1987). Politeness: some universal language usage. Cambridge, United Kingdom: Cambridge University Press.
Brown, J. J., & Reingen, P. (1987). Social ties and word-of-mouth referral behavior. Journal of Consumer Research, 14(3), 350-362.
26 psychological reactance: communicating threats to attitudinal freedom. In J. P. Dillard & M. W. Pfau (Eds.), The persuasion handbook: developments in theory and practice (pp. 213-232). Thousand Oaks, CA: Sage.
Burke, M., & Kraut, R. (2014). Growing closer on facebook: tie strength through site use. Paper presented at SIGCHI conference on human factors in computing systems, Toronto, Canada. doi: 10.1145/2556288.2557094
Chiou, J. S., & Cheng, C. (2003). Should a company have message boards on its web sites. Journal of Interactive Marketing, 17(1), 50-61.
Chu, S. C., & Kim, Y. (2011). Determinants of consumer engagement in electronic word-of- mouth (ewom) in social networking sites. International Journal of Advertising, 30(1), 47-75.
Cohen, E. (2003). Backpacking: diversity and change. Tourism and Cultural Change, 1(1), 95-110.
Coleman, J. (1990). Foundation of Social Theory. Cambridge, United Kingdom: Harvard University Press.
Dahlman, D., & Stafström, M. (2013). Female Swedish backpackers in vietnam: a hypotheses generating study on sexual health risks while travelling. Travel Medicine and
Infectious Disease, 11(1), 243-249.
Dillard, J. P., & Shen, L. (2005). On the nature of reactance and its role in persuasive health communication. Communication Monographs, 72(1), 144-168.
East, R., Hammond, K., & Lomax, W. (2008). Measuring the impact of positive and negative word of mouth on brand purchase probability. International Journal of Research in Marketing, 25(3), 215-224.
27 brand face-to-face but not on Facebook? How word-of-mouth on online social sites differs from traditional word-of-mouth. Journal of Consumer Psychology, 25(1), 120-128.
Evans, D. (2010). Social media marketing: the next generation of business engagement. Indianapolis, Indiana: Sybex.
Fournier, S., & Avery, J. (2011). The uninvited brand. Business Horizons, 54(3), 193-207. Fulgoni, G., & Lipsman, A. (2015). Digital word of mouth and its offline amplification: a
holistic approach to leveraging and amplifying all forms of wom. Journal of Advertising Research, 55(1), 18-21.
Gilbert, E., & Karahalios, K. (2009). Predicting tie strength with social media. Retrieved from https://experts.illinois.edu/en/publications/predicting-tie-strength-with-social-media
Giordano, C., & Giordano, C. (2011). Health professions students’ use of social media. Journal of Allied Health, 40(2), 78-81.
Higie, R. A., Feick, L. F., & Price, L. L. (1987). Types and amount of word-of-mouth communications about retailers. Journal of Retailing, 63(1), 260-278.
Huang, J. H., & Chen, Y. F. (2006). Herding in online product choice. Psychology and Marketing, 23(5), 413-428.
Hunter-Jones, P., Jeffs, A., & Smith, D. (2008). Backpacking your way into crisis. Journal of Travel & Tourism Marketing, 23(2), 237-247.
Hutton, G., & Fosdick, M. (2011). The globalization of social media: consumer relationships with brands evolve in digital space. Journal of Advertising Research, 51(1), 564-570. Jenkins, M., & Dragojevic, M. (2011). Explaining the process of resistance to persuasion: a
politeness theory based approach. Communication Research, 40(4), 559-590.
28 Internet and Higher Education, 4(1), 45-60.
Kaplan, A. M., & Haenlein, M. (2010). Users of the world, unite! The challenges and opportunities of social media. Business Horizons, 53(1), 59-68.
Katz, E., & Lazarsfeld, P. F. (1955). Personal influence: the part placed by people in the flow of mass communications. New York, NY: The Free Press.
Kelly, J. A., Lawrence, J. S., Diaz, Y. E., Stevenson, Y., Hauth, A. C., Brasfield, T. L., … Andrew, M. E. (1991). Hiv risk taking behavior reduction following intervention with key opinion leaders of population: an experimental analysis. American Journal of Public Health, 81(2), 168-171.
Knowles, E. S., & Linn, J. A. (2004). The importance of resistance to persuasion. In E. S. Knowles & J. A. Linn (Eds.), Resistance and persuasion (pp. 3-9). Mahwah, NJ: Lawrence Erlbaum Associates.
Koo, D. (2015). Impact of tie strength and experience on the effectiveness of online service recommendations. Electronic Commerce and Applications, 15(1), 38-51.
Lee, M., Rodgers, S., & Kim, M. (2009). Effects of valence and extremity of ewom on attitude toward the brand and website. Journal of Current Issues & Research in Advertising, 31(2), 1-11.
McWilliam, G. (2000). Building stronger brands through online communities. Slong Manage Rev, 41(3), 43-54.
Miller, C. H., Lane, L. T., Deatrick, L. M., Young, A. M., & Potts, K. A. (2007).
Psychological reactance and promotional health messages: the effects of controlling language, lexical concreteness and the restoration of freedom. Human Communication Research, 33(1), 219-240.
Murphy, L. (2001). Exploring social interactions of backpackers. Annals of Tourism Research, 28(1), 50-67.
29 Noort van, G., & Willemsen, L. M. (2011). Online damage control: the effects of proactive
versus reactive webcare interventions in consumer-generated and brand-generated platforms. Journal of Interactive Marketing, 26(1), 131-140.
O’Reilly, K., MacMillan, A., Mumuni, A. G., & Lancedorfer, K. M. (2016). Extending our understanding of ewom impact: the role of source credibility and message relevance. Journal of Internet Commerce, 15(2), 77-96.
Pfeffer, J., Zorbach, T., & Carley, K. M. (2013). Understanding online firestorms: negative word-of-mouth dynamics in social media networks. Journal of Marketing
Communications, 20(2), 117-128.
Reijmersdal van, E., Fransen, M. L., Noort van, G., Opree, S. J., Vandeberg, L., Reusch, S., Lieshout van, F., & Boerman, S. C. (2016). Effects of disclosing sponsored content in blogs: how to use of resistance strategies mediates effects on persuasion. American Behavioral Scientist, 60(12), 1458-1474.
Riet van ‘t, J., & Ruiter, R. A. C. (2013). Defensive reactions to health-promoting information: an overview and implications for future research. Health Psychology Review, 7(1), 104-136.
Rozin, P., & Royzman, E. (2001). Negativity bias, negativity dominance and contagion. Personality and Social Psychology Review, 5(1), 296-320.
Scherer, C. W., & Cho, H. (2003). A social network contagion theory of risk perception. Risk Anal, 23(2), 261-267.
Steffes, E. M., Burgee, L. E. (2009). Social ties and online word of mouth. Internet Research, 19(10), 42-59.
Sun, T., Youn, S., Wu, G., & Kuntaraporn, M. (2006). Online word-of-mouth (or mouse): an exploration of its antecedents and consequences. Journal of Computer-Mediated Communication, 11(4), 1104-1127.
30 Thompson, D. V., & Malaviya, P. (2013). Consumer-generated ads: does awareness of
advertising co-creation help or hurt persuasion? Journal of Marketing, 77(1), 33-47. Turner, J. C. (1991). Social influence. Bristol, United Kingdom: University Press.
Turner, V., & Turner, E. (1978). Image and Pilgrimage in Christian Culture. New York, NY: Columbia University Press.
Vance, K., Howe, W., & Dellavalle, R. P. (2009). Social internet sites as a source of public health information. Dermatologic Clinics, 27(2), 133-136.
Wang, T., & Yeh, R. K., Chen, C., & Tsydypov, Z. (2016). What drives electronic word-of- mouth on social networking sites? Perspectives of social capital and
self-determination. Telematics and Information, 33(1), 1034-1047.
Zhang, N. (2014). Web-based backpacking communities and online activism in china: movement without marching. China Information, 28(2), 276-296.
Zhao, M., & Xie, J. (2011). Effects of social and temporal distance on consumers’ responses to peer recommendations. Journal of Marketing Research, 48(3), 486-496.
Appendix
Table 1. “tie strength” & “risk perception” (not enough N for weak tie in Univariate Anova, was not carried out)
M SD
Normal .50 .71
Weak tie / /
31 Table 2. “tie strength” & “SWOM engagement”
M SD
Normal 1.00 1.41
Weak tie 2.00 /
Strong tie 2.00 .00
Table 3. “type of language” & “risk perception”
M SD
Unforceful language 1.00 .00
Demeaning language .50 .71
Controlling language 1.00 .00
Table 4. “valence of reactions” & “risk perception”
M SD
Neutral valence .80 .45
Positive valence 1.00 .00
32 Table 5. Descriptives N N in experiment Beginning 652 Excluded 279 Randomized 595 Dropped out 277 Analyzed 373
33 Table 6. Descriptives for Tie Strength, Valence of Reactions and Type of Language on Risk Perception for gender.
Risk perception
Total Tie Strength Valence of
reactions
Type of Language Strong tie Weak tie Positive
valence Negative valence Controlling language Demeaning language Unforceful language Sample size 373 131 101 55 Sample size male 25 2 20 1 6 / 2 Sample size Female 95 5 32 29 38 2 7
34 Table 7. Descriptives for Tie Strength, Valence of Reactions and Type of Language on SWOM Engagement for gender.
SWOM Engagement
Total Tie Strength Valence of
reactions
Type of Language Strong tie Weak tie Positive
valence Negative valence Controlling language Demeaning language Unforceful language Sample size 373 105 93 67 Sample size male 21 4 7 10 11 / / Sample size Female 63 8 23 30 45 6 5
35 Table 8. Age.
Conditions M SD
1. Strong tie, positive valence, demeaning language
24 /
2. Strong tie, positive valence, unforceful language
25 /
3. Strong tie, positive valence, controlling language
25,33 3,67
4. Strong tie, negative valence, demeaning language
24 /
5. Strong tie, negative valence unforceful language
/ /
6. Strong tie, negative valence, controlling language
36 7. Weak tie, positive valence, demeaning
language
/ /
8. Weak tie, positive valence, unforceful language
/ /
9. Weak tie, positive valence, controlling language
/ /
10. Weak tie, negative valence, demeaning language
/ /
11. Weak tie, negative valence, unforceful language
/ /
12. Weak tie, negative valence, controlling language
