A mixed methods study of the adolescent approach to online risks.
Research Master’s program Communication Science Specialization Youth & Media
Master Thesis
Eva Hollander 5800544
January 6, 2014
Abstract
Research on risk behavior online has mainly focused on quantitative data about the general behavior of adolescents online. In addition to the knowledge on general online behavior of adolescents, this study focuses on the attitudes and knowledge of online risk behavior from the perspective of the adolescent.This research was conducted with a mixed methods approach.
Study 1 was conducted with focus groups among 26 adolescents in the fourth grade (15-17 years old). In these focus groups different topics concerning online behavior (privacy online, risks online and influences on risk behavior) were discussed. Overall a paradox was found between the knowledge and behavior of the participants. Most of the adolescents were positive of their own online behavior, but several mentioned behavior that is perceived as risky, such as contacting strangers online.
Study 2 was conducted with a content analysis of Facebook profiles. Results indicate a support of the paradox that was found in the focus groups. The information provided on a profile differs between the privacy levels. This is expected since private profiles entail little information about the profile owner. But, several results indicate that the risky behavior online is not in the amount of general information given within a profile, but in the self- portrayal in pictures.
Close Encounters of the Adolescent Kind.
Everyday threats are expressed through social media, children are being bullied online, pedophiles search for victims on the Internet and the reach of sexual predators has extended with the emergence of 'sexting' (Christofides, Muise & Desmarais, 2012). The act of sexting has been defined as ‘the practice of sending sexually explicit images or text through mobile phones or via internet applications’ (Hasinoff, 2013, p.449). There are a lot of ways in which criminals can take advantage of information they find online, e.g. conversations and pictures can be spread to any number of individuals or personal information may be stolen for identity fraud (Whitty &
Joinson, 2008; Saeri, Ogilvie, La Macchia, Smith & Louis, 2014). These are consequences of the violation of privacy, which is defined by Yen & Tu (2011) as 'the ability to maintain personal information and social acts from the public' (p.236).
According to the study of Newcom (Boekee, Engels & Van der Veer, 2014) , 4 out of 10 Dutch youngsters (participants of this particular study are 15-19 years old) are concerned about privacy issues online and the control over personal information. By default, the privacy setting of personal information on a Facebook profile is 'friends only', which means that the information is only available to the individuals online friends group (in this study referred to as the ' private' setting). But, an individual can also choose to share the information of their profile with a wider audience, such as 'friends of friends' or 'public' which makes the profile accessible to anyone.
In addition, many studies (Baumgartner, Valkenburg & Peter, 2010; Christofides et al., 2012; Delevi & Weisskirch, 2013; Doornwaard, Moreno, Van den Eijnden, Vanwesenbeeck & Ter Bogt, 2014) have looked at the occurrence rate of these risks among various age groups. Even though there is a lot of statistical data available about the use of these platforms among teenagers and adolescence, qualitative research has remained scarce. This thesis will show that, in addition
to this statistical data, it is important to talk to young people about their attitudes towards and knowledge of online risks.
This study revolves around adolescents, which is a term for young people between 10- 22 years old. This study is conducted on high schools, resulting in an age group of 14-19 years old. During adolescence the development from child to adult is at full speed and the development focuses on three levels; physical, cognitive, and social- emotional development. As will be discussed in the next chapter, these developmental stages are not always on the same level. This is why adolescence is termed as period of risks and opportunities (Nelis & Van Sark, 2014).
The study has been inspired by the interests and spearpoints of the Kennemerland police force, which is responsible for several communities (small villages to cities). Its communication department is interested in the behavior of young people online, as part of its duties to prevent crime and inform the public. The emergence of social media requires a completely new focus, especially with social environments of high schools where social networks of adolescents are established. Both focus groups and content analysis are employed in this study, which allows for a comparison between perceptions and knowledge about online risks, and actual online behavior. This mixed methods approach will be the foundation of the answer to the following research questions: What is the attitude towards and knowledge of online risk behavior among
Theoretical framework
In the Netherlands almost all (930.000 ) youngsters (15-19 yrs old) use Facebook, with 1 745.000 of them using this social network site (SNS) on a daily basis (Boekee et al., 2014). Adolescents represent the group that uses social networking sites more frequently than other ages (Christofides, Muise & Desmarais, 2011). This online environment challenges them to
experiment with identity and skill development through various forms of online activities such as networking, chatting in web communities and online gaming. Even though the Internet provides interesting opportunities for youngsters to entertain, educate and develop themselves, it also provides potential risks, which include cyber bullying and sexual harassment (Keipi & Oksanen, 2014). Social network sites are based on profiles, which Danah boyd (2007) describes as 'a form of individual homepage, which offers a description of each member' (p. 123). The concept of privacy has been a contentious issue ever since the Internet expanded. Since the information on the Internet can be kept forever, searching for information has become very easy and can be assessed without the person of interest knowing (Saeri et al., 2014). According to Faisal and Alsumait (2011), gender is one of the most important factors that affect online privacy concerns. In addition, Saeri et al (2014) present results indicating that women have higher privacy
protection behavior than men. This results in a first hypothesis that states:
H1: Gender has an influence on privacy protection behavior (woman have greater privacy protection settings (= more private) than men.
This study contributes to the clarification of this concern by applying a mixed methods analysis of the attitudes and behavior towards online risks among Dutch adolescents. The theoretical background will clarify some of the more frequently identified online risks in general, as well as
According to Newcom (Boekee et al., 2014), there are 992.851 individuals from the age 15-19 living in the 1
the development of adolescents, the role of risks and trust, awareness in adolescence and behavior online of this particular age group.
Who are adolescents?
Adolescence is a period of fast change, which is 'dramatically crystallized in the flood of hormonal activity and rapid physiological development that constitutes puberty’ (Crosnoe & Johnson, 2011; p. 440). Developments of both body and mind call for a quest for personality. During this period individuals are perceived as neither child nor adult. In addition, the social orientation shifts from parents to peers which is expressed by spending more time with friends than with family members (Baumgartner et al., 2010, Jordán-Conde, Mennecke & Townsend, 2013). Individuals start to experiment with their identity before they commit to adulthood. This struggle is countered by comparing their identity to others who act as a 'mirror to the
self' (Jordán-Conde et al., 2013; p. 358). The Internet functions as a major source of information for adolescents as well as a communication spot where adolescents spend time exchanging feelings and thoughts (Jordán- Conde et al., 2013). Jordán- Conde et al. (2013) describe the concept of 'disclosiveness' as the 'tendency to share intimate information that is highly personal with strangers' (p. 359). The sharing of intimate information is driven by the feeling of 'being there' which interacts with the personality crisis of adolescents and the experimentation of
identities and intimacy (Jordán-Conde et al., 2013). Sung and Mayer (2012) describe the sense of presence in an online learning environment as: 'whether participants feel they are interacting with real people when they are online’ (p. 1738). According to their theory, this feeling of 'being there' is related to five facets; social respect, social sharing, being open minded, social identity, and intimacy. This is a key aspect to understanding online communication and refers to the degree to which a person in perceived as a 'real person' in online communication. The application of this theory is slightly different in the online risk debate that is outlined in this study. However, three
of the facets clearly apply here; social sharing, social identity and intimacy. First of all, the sharing of personal information builds social relationship. Sharing information and discovering common grounds elevates the feeling of being comfortable with another person. In addition, the social identity as it is represented by a person’s online profile needs to be recognizedand
accepted by the social environment.This enhances the level of intimacy between two individuals and encourages to share personal stories and experiences (Sung & Mayer, 2012). Yen and Tu (2011) add the dimension of privacy as an important aspect of social presence. This is defined as 'the sense of being secluded from the view of others' (p.222). This creates an environment in which there is less concern for sharing of information.
The Internet can be seen a platform where an online identity can be constructed and has changed the traditional process of identity construction (Zhao, Grasmuck & Martin, 2008). It provides an environment detached from the body with abilities to interact with one another without revealing any physical characteristics. This allows for a more free and creative
environment to engage in identity exploration. A SNS like Facebook can function as a platform for exploring identities. The online personality can then act as an extension of the real life self that functions as a way to explore identities (Jordán-Conde et al., 2013; Lee, 2014).
Anonymity can be maintained online through withholding personal information. According to Zhao et al. (2008), the combination between the disembodiment and anonymity provides a technical environment in which a new process of identity formation can emerge. The process of constructing a Facebook profile allows the profile owners to present a selected self- image and therefore engage in 'impression management’. For the profile owner this provides the
opportunity to create a desirable, or even misleading, impression. For the viewer of this profile, this provides the challenge of determining which information gives a true reflection of the offline personality of the individual behind the profile (Hall, Pennington & Lueders, 2013). Social
network sites therefore provide an ideal platform of impression management (Krämer & Winter, 2008). Faisal and Alsumait (2011) agree with this statement and mention that one major aspect of discovering identity online is the portrayal of oneself online by the amount of information the participants disclose. These profile items are considered sensitive, because they reveal the true identity of the individual. Specifically, the public statement of the relationship status can have interpersonal and social implications, according to Fox, Warber and Makstaller (2013). The findings of their study reveal that Facebook can act as a way to reduce uncertainty in the first stages of a relationship. Communicating our relationship status is not a new phenomenon, think of wearing a wedding ring, but on SNS it is easier and faster to provide information about romantic relationships to a larger network (Fox, Warber & Makstaller, 2013).
The role of risk and trust
With these developments, adolescents learn to better understand the perspective of others and shift between the perspective of self and others, which is a characteristic that is associated with prosocial behavior. A key element of this behavior is trust (Van den Bos, Westenberg, Van Dijk & Crone, 2010). Trust increases from childhood to adulthood and research indicates that trust is higher in adulthood than in childhood and adolescence (Sutter & Kocher, 2007).
Saeri et al. (2014) discuss perceived trust in combination with perceived risk as two important affective variables on online risk behavior. The perception of risk is an important factor in social judgment and alerts us to potential negative consequences of our actions. The fading lines between online and offline environments, and the illusion of intimacy on the Internet, can reduce the role of risk perceptions online (Saeri et al. 2014). According to Saeri et al. (2014), trust is based on a positive relation of being vulnerable to others. A certain way to create mutual trust is through self-disclosure, or the sharing of personal and private information. By disclosing personal information an individual is perceived as more trustworthy.
However, if a profile is set to 'friends of friends', the exhibition of personal information can extend to a much wider audience than realized.
The study by Saeri et al. (2014) has found that participants who perceive greater risks and have lower trust in others would be more motivated to protect their online privacy. However, it remains questionable whether motivation alone is enough, as there was no association found between the actual behavior and perceived risk. This may possibly be explained by the impact of social comparison, where peer comparison results in information that influences risk perceptions (Saeri et al., 2014). The social comparison theory of Leon Festinger (1954) explains that
individuals compare themselves with others when they are confronted with information about other individuals. People can either feel negative or positive about themselves as a result of comparison with others, but the main goal for this behavior is to enhance one’s self- esteem (Lee, 2014). Perceived trust was mostly associated to the attitudes towards online privacy protection. Those who were more positive towards protecting their privacy online had lower trust in other Facebook-users. In addition, gender differences were found in perceived trust and the correlated online behavior. Woman were less trusting of others online then men, and displayed more privacy protection behavior over their personal information. And a predicting effect of age was found in this study, where older participants (adults) had more intentions to protect their privacy online than younger participants (adolescents) (Saeri et al., 2014).
Related to these concepts of trust and risks is the notion of perceived invulnerability, which is assumed to rise due to the cognitive development during adolescence. Since the adolescents’ brain is not fully developed yet, they are assumed to be particularly susceptible to the believe that one is unique and invulnerable (Baumgartner et al., 2010). According to
Baumgartner et al. (2010), the perceived invulnerability is closely related to low perceptions of risk. However, even if adolescents have high risk perceptions they may still feel personally
invulnerable. This might be explained by the neurological stage of adolescence where the prefrontal cortex is not fully developed yet (Nelis & Van Sark, 2014). This part of the brain is responsible for abilities like controlling impulses, judgment, solving problems and social
behavior. As a result, adolescents often feel quite independent and mature, but behave otherwise and have not developed complex skills such as long- term thinking. In addition to the cognitive development, the physical and social- emotional stage is often on a different level. While the sexual maturation is completed, the social- emotional and cognitive development is still undergoing major changes. The production of hormones is at full speed and the need for kicks resulting in an adrenaline explosion is high (Nelis & Van Sark, 2014). The cognitive dissonance theory of Festinger (1962) is also interesting in this context. Leon Festinger (1962) suggests that we possess an inner drive to keep our attitudes and beliefs in balance and avoid disturbances of this equilibrium (which he calls cognitive dissonance). We strive for consistency within our attitudes and opinions. This consistency also applies to what a person believes and what he does. What stands out then, are behaviors that are opposed to this consistency. The inner drive to maintain the balance can be strong, resulting in irrational behavior in order to maintain the cognitive balance. Resulting in attempts to rationalize the inconsistent behavior. If these attempts fail and the presence of inconsistencies is not rationalized, psychological discomfort appears (Festinger, 1962).
Even though the adolescents in the study by Baumgartner et al. (2010) have high perceptions of risks, they sometimes still feel invulnerable. This might be explained by the cognitive dissonance that arises when one acknowledges the risks, but still engages in risky behavior online. They would then rationalize the personal circumstances that apply to them, which make them believe they are untouchable. These adolescents would then deny potentially negative consequences and focus on the benefits in order to justify their behavior.
Identifying risky behavior
Generally, risk behavior can be defined as every behavior that involves potential negative consequences. With this definition, many online behaviors can be perceived as risky
(Baumgartner et al., 2010), such as phishing (the act of obtaining personal information without the permission of the individual concerned for illegal purposes), downloading of illegal content and hacking. Part of these behaviors can be defined as risky sexual behavior; 'the exchange of intimate, sexually insinuating information, or material with someone exclusively known
online' (Baumgartner et al., 2010; p.1227). One such risky sexual behavior that has been widely discussed since the development of picture- based social media such as Snapchat and Instagram is 'sexting', which may be defined as ‘the practice of sending sexually explicit images or text through mobile phones or via internet applications’ (Hasinoff, 2013; p. 449). One major risk of sexting is that it is often spread among a wider audience than that the sender intended. The impact on the participants of one single act of sexting can range from damaging the reputation of the producer, cyberbullying or other forms of peer aggression (Van Ouytsel, Walrave & Van Gool, 2014).
Another risky behavior that is categorized is the disclosure of personal information. One potential negative consequence of online self- disclosure is the use of personal information by others with malicious intent. Christofides, Muise & Desmarais (2011) found a difference in amount of self- disclosure online between adolescents and adults. One explanation for the
difference in privacy behavior between adolescents and adults, is that the development of privacy sense is not fully developed yet. Youth does care about their privacy, especially when it is about privacy from their parents (Christofides et al., 2011). But personal information such as age, religion and sexual preference are not considered information that needs to be private on social networking sites (Livingstone, 2008). Another explanation follows from the social cognitive
theory of Bandura (2001). This theory provides a theoretical framework for understanding the processes of human behavior (Wu & Chen, 2013). In line with the social cognitive theory, when users of Facebook see their friends disclosing information, they will adopt this behavior and disclose more information as well (Christofides et al., 2011). The concerns of disclosing personal information online lead to the following hypotheses:
H2a: Gender has an influence on self- disclosure in general.
H2b: Privacy settings have an influence on self- disclosure in general.
H3a: Gender has an influence on the listing of the current relationship status.
H3b: Privacy settings have an influence on the listing of the current relationship status. H4a: Gender has an influence on the amount of personal information in status updates. H4b: Privacy settings have an influence on the amount of personal information in status updates.
The worries about social networking sites mostly involve inappropriate behaviors on SNS. This study focuses on the pictures which are publicly visible on the Facebook- profiles of the sample. The classification consists of alcohol expressions, nudity and sexual behavior in pictures.
Alcohol expressions and sexual risk behavior
Worries about portrayals of alcohol use in the media is not a new phenomenon (Anderson, De Bruijn, Angus, Gordon & Hastings, 2009; Van Hoof, De Jong, Fennis & Gosselt, 2009). As Beullens and Schepers (2013) state that research on alcohol portrayals on SNS is important for several reasons. First of all, the majority of youngsters use SNS on a daily basis. In addition, people are affected by the social cognitive theory in which states that individuals are influenced by the behavior of others. On Facebook, users receive feedback on their own behavior and are exposed to the behavior (and related feedback) of others. This is important because several studies (Beullens and Schepers, 2013; Cail & LaBrie, 2010; Teunissen, Spijkerman, Prinstein,
Cohen, Engels, & Scholte, 2012; Jones & Magee, 2014) have found perceived peer drinking to be a significant predictor of alcohol use. In addition, Moreno, Briner, Williams, Brockman, Walker & Christakis (2010) have found a significant difference between the frequency of alcohol portrayals for males (54.2%) and females (45.8%). Therefore, it would be interesting to apply these findings to this study, resulting in the following hypotheses:
H5a: Gender has an influence on the level of the alcohol expressions in pictures.
H5b: Privacy settings have an influence on the level of the alcohol expressions in pictures. In addition, the level of nudity and portrayal of sexual behavior in pictures are other forms of inappropriate online behavior. According to Hew (2011), portraying such behavior may jeopardize chances in the future if future employers search for information about candidates online. Next to the risks for future activities, is the risk of other people use the information they find online in a harmful way (Christofides et al., 2012). Therefore, the final hypotheses that will be tested in this study:
H6a: Gender has an influence on the extent of nudity in pictures.
H6b: Privacy settings have an influence on the extent of nudity in pictures. H7a: Gender has an influence on the extent of sexual behavior in pictures.
H7b: Privacy settings have an influence on the extent of sexual behavior in pictures. Method
A mixed-method approach was conducted among adolescents in the age group of 15-18 years by means of focus groups and content analyses. The first study consists out of qualitative focus groups on the topic 'online risks'. The choice for this approach results from the gap in existing research on risk behavior online among adolescents. Even though much quantitative data is available about the knowledge and behavior online and the understanding of risk consequences among adolescence, less research is conducted on the perceptions and attitudes
according to the adolescents themselves. In addition, group behavior during the focus group discussion can provide information about the dynamics in different groups when talking about sensitive subjects. These details about the dynamics of a group will not surface in quantitative research or individual interviews. These focus groups have been conducted between the 10th of April and the 4th of June 2014. All focus groups took place in Haarlem and lasted approximately 45 minutes (the duration of one school lesson).
The second study was conducted with a quantitative content analysis of 154 Facebook profiles. As will be discussed in more detail in the 'Results' chapter, most of the participants of the focus groups indicated that they were well aware of the privacy settings of their online profiles, but when we discussed the different aspects of privacy of a profile (such as the number of Facebook friends) signs of a discrepancy between knowledge and behavior started to appear. In order to discuss this result in more detail, the focus in the second study was on the amount and nature of information that is placed on Facebook profiles.
Study 1- Focus groups Sample
The sampling procedure of the first study was purposive. The research was executed at high schools, which were contacted by e-mail. This e-mail provided a summary of the purpose of the research and gave an overview of the structure of the focus groups. In each case, the contact person offered to recruit the participants in his classes prior to the focus groups. He/ she then sent the passive consent form to the parents of the concerning students (Appendix A (in Dutch)).The voluntary nature of the study was made clear and parents or participants were also able to withdraw the participation in the focus groups afterwards. In total 26 adolescents participated in the focus groups (12 females and 14 males), none of the participants/parents refused
participation or withdrew the participation after the focus groups took place. The participants only received drinks and candy during the interviews, no monetary reward was provided.
Procedure
The participants all gave permission to record the interview. An iPhone was used for recording audio. These recordings were transcribed afterwards and then coded with Atlas.ti (qualitative data analysis tool). The interviews were held in Dutch, and transcribed in Dutch as well (Appendix F). The focus groups were conducted based on an interview guide. A short
introduction explained the central focus of the thesis, the importance of their attendance and the main topic of the interview(online risks behavior). In addition, the anonymity of the study was emphasized; only fist names would be used, and could be altered if wished. Next, the
participants were asked to note down their first names and their age on a nameplate. Even though the focus groups discussions were anonymous, the names were required to transcribe the group interviews. Afterwards the names were altered. The interview started with some questions about their use of social networking sites and proceeded with an association game about their
knowledge of online risks, followed by questions about their behavior towards these risks and the influences on their awareness of online risks.
The interview guide
The interview guide provided a basic structure for the focus groups (Appendix B (in Dutch)). The interview started with an introduction where the research was briefly explained and
information about the processing of the interviews was provided. Furthermore, an emphasis was made on the informal setting to provide a comfortable atmosphere in which the participants were free to speak their minds regarding the topics. The interview guide was divided up into four topics: (1) the use of social networking sites, (2) what are the risks?, (3) handling online risks, and (4) influence on engagement withonline risk behavior.
The first topic was the use of social networking sites. This was a preludingtopic which was included to set a comfortable atmosphere. Furthermore, the questions regarding this topic were aimed to establish the use of social networking sites among the participants, since this was a requirement for the rest of the interview questions. The purpose of these questions was to understand which social networking sites the participants use, and which ones are the most popular.
The second topic of the interview guide was ‘what are the risks?’. This topic involved an association game. The participants were asked to individually write down as many keywords that they associated with online risks within three minutes. Each keyword was noted on a new post- it. These keywords were then discussed in the group when they were all finished. First the keywords were mentioned individually and then questions were asked about the awareness of these risks, and to what extend the topic is discussed among peers. This topic was aimed at the perception of online risks among the participants.
The third subject was ‘handling online risks’. The previous topics established the understanding of online risks among the participants. This topic was aimed at the attitudes towards online risks that are encountered. The participants were asked to indicate how they act when they come across online risks. In addition, this topic continued by discussing why teenagers would talk to strangers online.
The fourth and last subject was the ´influence on engagement with online risks?’ This topic focused on the differences between teenagers who engage in online risks, and teenagers who don’t. The aim was to talk about why this difference exist and what can affect a teenager in his or her choices to portray risky behavior. Furthermore, the effectiveness of information campaigns was shortly discussed.
The interview ends with a final question about what they thought could help in making teenagers and adolescents more aware of online risks. This was an open question with room for discussion about the participants’ opinion on a successful campaign and to provide an
opportunity to share additional thoughts. Study 2- Content analyses
Sample
A content analysis was conducted with 154 Facebook profiles. A snowball sampling method was chosen in order to sample a comparable high school- group by searching on educational level groups of last school year (2013-3014) or this school year (2014-2015). This resulted in six groups based on educational level, from each group one profile was selected to start the snowball sampling. While this resulted in a comparable sample, the method used was too specifically targeted at educational level with exclusion of this variable in the analysis as a result. The
reliability of the findings would otherwise be in jeopardy, since we would then generalize results from a sample that is too specific.
The sample consisted of 89 profiles (57,8%) belonging to females and 65 profiles (42,2%) belonging to males. These profiles were first coded on privacy settings as specified in the
codebook (Appendix C). The privacy settings were separated at the general information, photo’s, interests and timeline of the profile.
Procedure
The profiles were sampled and analyzed during two weeks in August 2014. A codebook was designed beforehand (Appendix C). No informed consent was needed for this content analysis, since the level of privacy settings already determined the view ability by others who visited a certain profile. The analysis of the profiles was conducted by one researcher. Anonymity was guaranteed by giving an item number to each profile in stead of the participants’ name. In order
to be able to return to a certain profile the item numbers were tracked in a private document of the researcher in which the item number was combined with the link of the concerned profile. Since this study is anonymous, this document is not attached.
The content analysis was conducted with a codebook that was designed on Qualtrics.com. First, the analysis focused on the privacy settings of four parts of the profile (general
information, photo’s, interests, and the timeline). Then several general questions about the
participant were answered, such as the number of Facebook friends, and age of the participant. In addition, several topics about demographics were answered with known/ unknown answering scales, such as ' residency', 'sex interest in..?', and 'current relationship status'. All of these topics were only answered as 'known' when this was explicitly indicated in the general information about the participant. Next, the photos on each profile were analyzed on the topics that were found in the profile. Several questions focused on specific behavior and whether or not this behavior occurred in the pictures on the profile, such as sexual behavior and alcohol references. Finally, the timeline was analyzed with questions about the information that was found within the status updates. Each content analysis took 5 to 15 minutes to complete depending from the privacy level of the different parts of the profile, for instance a profile that is publicly visible entails more information than a profile that is private.
Measures
Information control. The first questions concerned the privacy settings that were assigned to the different parts of the profile; general information, photo’s, interests (such as movies, books and music artists), and timeline. The answering scale was based on the options provided by Facebook (private or unknown= only friends are able to view the profile information, partially private= friends of friends are able to view the profile information , and public= everyone is able to view the profile information).
In addition to the privacy settings of the profiles, the number of Facebook friends were indicated with an answering scale from unknown, 100 or less up to more than 500. This question was part of the information control measure since the focus group results indicated that
participants can be aware of the privacy settings, but their awareness can be abolished by high numbers (from 300 and more) of Facebook friends. Especially when the privacy settings are set at ' partially private', which means that all friends of the connections are able to view a profile. This can quickly result in thousands of individuals who are able to look at personal information.
Information disclosure. The information disclosure concerns the personal information of an individual that is listed on a profile. These questions are derived from the study of Zhao, Grasmuck & Martin (2008) with personal information topics, such as 'Added a ‘about me’ description', 'Sex interested in..?', and 'Current relationship status' which are answered on a known/unknown- answering scale. In addition, the coding of the information provided by the status updates was based on the research of Bernstein et al. (2010) with multiple answers possible (e.g. personal information, news, and change of profile and/or cover picture).
Risky behavior. This topic was coded with several questions in the content analyses. The first one was based on the research of Fournier & Clark (2011) and concerned the
communication about alcohol related activities (e.g. parties) which was measured on a 7-point Likert scale ranging from 'never' to 'daily'. A second question concerning alcohol was derived from Atkinson et al. (2011) and concerned the explicit references to alcohol ranging from 'no references' to 'active drinking act' (4-point scale). A distinction was made between alcohol references on the profile and alcohol references in pictures both using the same answering scale. The second behavior that was coded was sexual behavior. This question concerned sexual behavior in pictures and ranged from 'none' to 'very explicit' (4-point scale). And the last
question that measured risk behavior was about the extent of nudity in pictures which was measured on a 'none' to 'very explicit' (4- point scale) answering scale as well.
Results Study 1 Focus groups
The discussion of the results of the focus group interviews are guided by the concept indicator model. This model consists out of two concepts, which are divided into several
dimensions that define the concept. The main concepts of this thesis are the awareness of online risks, and the apparent paradox between this awareness and the online behavior. These are based on several dimensions. The concept awareness consists of ‘influences on awareness and
behavior’, ‘perceptions on online risks’, and ‘decreasing risky events online’. The concept behavior consists of ‘uses of social media’, ‘purpose of social media use’, ‘perception of online risks’, ‘behavior towards online risks’, and ‘attitudes during focus groups’. These dimensions are derived from the focus group discussions. The concept indicator model is to be found in
Appendix D. Awareness
Influences on awareness. This first dimension of the concept 'awareness' defines the influences on the awareness of online risks. Positive influences on online risk awareness were parents and news items or television programs. Respondents mentioned the age of 12 as the most affective and logical starting point of influence on awareness. This is because, according to the respondents, around the age of 12 the use of Internet and SNS intensifies. At this point, parents should talk to their children about online standards, according to several participants. Comments on this influence reappeared several times in different groups:
Wouter: Parents have the most influence on direct dangers, such as information about your house. For example, a post about going on a holiday and then someone will break into your house. 2
In addition, the matters that concern parents the most have a practical nature. They might be influenced by media coverage on online risks that often involve the matters that are discussed by the participants. In addition, they approach these adolescents as young people that need to be educated about the practical parts of life. Furthermore, they will try to keep their children safe by giving them obvious examples of the consequences of online self- disclosure.
Roos: My mom always warns me that I should not post that we are on a holiday.
This example provided by Roos appears several times. The participants discuss the worries of directly dangerous matters that consume the parents. They emphasize that you need to be cautious with personal information, such as your home address and posting updates of your vacation. According to the parents, this might attract burglars.
Not all participants agreed with the positive influence of parents. Some of the
respondents discuss the difference in generation. They grew up with technological devices such as computers and describe themselves as ‘digital natives’ who speak the language of the Internet fluently. This is in stark opposition to their parents who are labelled as ‘computer nerd zero’ and have difficulties with the basics of a computer.
Patrick: my dad was computer nerd zero. I even had to show him how to make a folder on the computer, so he can’t do much on the computer anyway. My mom had some skills, but I don’t meddle in their Internet use.
Or they claimed that their parents had no influence at all:
Chris: I am careful on the Internet, but they [parents] have nothing to say about it actually. It is not her [mother] problem.
These participants distanced themselves from the Internet use of their parents and made it clear that their awareness online resulted from their own knowledge. Adolescents are notorious for wanting to make their own choices, even though their brain is not fully developed yet. This often results in choices that are not well thought through. The actions of Chris’ mother clearly work counterproductive with Chris engaging in obvious adolescents’ behavior: rebellion against the parents’ view (Nelis & Van Sark, 2014).
Perceptions on online risks. When discussing the associations with online risks with the participants several topics were recurrent. Most prominent was the Catfish- association that was mentioned several times by the male as well as female participants. This refers to fake accounts on social networking sites, with the intention to make contact with other people online.
In addition, the associations with online risks forwarded prominent differences between the boy and girl- groups. The boys mentioned quotes which can be used against you, because anything you put on the Internet can still be retrieved after a long time.
Thijs: Like asking on the Facebook- page of some party ‘where can I buy the pills?’
This risk increases when you are applying for a job:
Wouter: if you portray yourself as vulnerable on the Internet, and say things which are intended for friends, than you can sometimes retrieve these quotes through Google, and an employer might not hire you.
Even though the prefrontal cortex is not fully developed yet, the answer given by Wouter indicates that young people can be aware of the long- term consequences, especially with practical matters that are controllable and involve daily live.
important risks being viruses and so- called data- streaming (hacking into computer systems and stealing bank details). While none of the female respondents mentioned ´virus´ this online risk was coded thirteen times in the male focus groups.
Erik: My computer completely crashed from a virus once. I don’t know what happened.
These findings indicate a gender difference in Internet use, where boys spent their time online playing games and girls chat or shop online. In addition, this could also indicate a process that is explained in the theoretical framework, the quest for personality. It maybe possible that boys focus more on the technical, masculine risks of the internet than more vulnerable risks in which a victim and perpetrator is involved, such as with cyberbullying. This would then confirm a strong divide between the sexes and illustrate a normalized sexism in which a strict gender divide has saturated the adolescents lives.
These associations with online risks were based on the keywords that participants wrote down. Several participants struggled to come up with their own associations and copied the associations from other, even though it was explained that the associations should be noted down individually.
Geert: O, o, I don’t think about anything, that is the problem. *Laughter*
Patrick: Uhh uhh, this is hard, uhhm. Robin: I’m really not good at this. Patrick: Whoever has more Robin Robin: Shit! I give up.
The girl participants discussed different topics as most important online risks. Cyberbullying, fake accounts and privacy were mentioned most often. In addition, other striking associations were mentioned:
Jojet: Intimidation (sexual), spam, stalkers, loss of concentration. You are checking your social media often on the go. And young people are afraid that they will miss out on something. Eva: And what do you mean with intimidation (sexual)?
Jojet: Nude pictures can be spread really quickly. The way to spread these pictures is much more easy.
These topics were recurrent within the female focus group discussions, but were not mentioned by the male participants. Again, the examples mentioned by Jojet illustrate the more vulnerable, and negative, part of the Internet.
Decreasing risky events online. As a last topic for the concept awareness the participants of the focus groups discussed the possibilities to reduce online risk behavior. Most of the
adolescents agreed that the information provided about these online dangers should be improved and extended. It was often mentioned that a school environment would work best, for example by organizing a session about this topic for all students.
Pim: Actually, I do think that it can make a big difference. Because ideas are contagious. Once it is in your head, it can result in making different choice than you otherwise would.
The opinions were divided at the point whether or not these gatherings should be mandatory. Furthermore, the participants indicated that it would be a good idea to invite the parents to these gatherings as well, because they are most influential in raising the awareness outside the school.
Sander: Yes if you want to organize an information gathering you should make it mandatory[..].
Roy: Yes then you gather a whole group at once if you make it mandatory. It should be for the parents as well, because they will talk about this at home.
shown by vivid examples that match their phase in life. This is possibly the result of the intrinsic motivation processes of adolescents. One of the triggers to intrinsic motivation to talk and think about online risks, could be the feeling of the existence of a relation between individual action and long- term consequences. Even though this ability is not yet fully developed in adolescence, handing them personal negative long- term consequences on a plate, could trigger a motivation to think about these consequences.
Erik: I do think information gatherings can work if you organize this in the beginning of the second year of high school and you invite all students and parents. If you take an example of someone who has been crushed by something that has destroyed his live, than you will definitely be more aware of what you put on the Internet.
This seems to be opposite of the policy of the schools in the area where this study was
conducted. The students overall indicated that they did not receive any education or training in how to handle online risks. This was confirmed by the local police force who explained that the schools acknowledge the dangers online, but are hesitant with putting this topic in a spotlight. Reason for this is that they don’t want to give parents the idea that these events occur on their school.
Apart from these information gatherings the discussion focused on advertisement clips on social media. A recurring example is an advertisement in Belgium that is explained in the following quote:
Alex: there was this advertisement on Facebook about exceeding speed limits. It was this advertisement in Belgium. People that were often exceeding speed limits were led into a big building and into a room with a coffin. They were asked to sit down and told that they were invited to their own funeral. Next they were told how they had died and that you should not exceed speed limits because this could be the consequence.
According to the participants, the Dutch information advertisements are too exaggerated. They do not believe these events will happen to them, because they are portrayed unrealistically in the clips. The advertisements should be more focused on their daily lives and adjusted to their age and interests. If the advertisement or information clips are too exaggerated, they would feel like it will not be a consequence that will apply to them because they can’t fit the scenes in their own daily lives.
Behavior
The digital environment (uses of social media). All the participants indicated that the use of a computer, tablet or mobile phone was part of their daily routine. Most of the participants also confirmed that the use of social networking sites is an important activity within these forms of communication, with Facebook as the frontrunner followed by Instagram. In addition,
Whatsapp was frequently mentioned as well, but several participants were unsure if they would consider this application a form of social media. Nevertheless, these smartphones are popular among young people, and present a constant access to Internet, without requiring WiFi.
Many of the respondents talked about the decline in the use of their Twitter-account. While they were frequent Twitter users only a year ago - some even spoke about tweet counts up to 40.000 - the network site has now lost its attractiveness. The respondents claim this is the result of an increasing scope of social networking sites and the information posted on Twitter was good-for-nothing.
Mirthe: It depends on your age, when I was younger I thought it was fun to put everything on Twitter. But actually, no one gives a damn.
The reaction of Mirthe indicates that the use of Twitter is assumed to belong to a younger age and little thought is put into the information that is posted on the social network. It could also be
the result of peer influences; if friends or classmates use Twitter, the rest of the peers will be likely to follow and the same goes for its decline. They would then conform to the behavior of their peers, which is a characteristic of early adolescence. Around 14-16 years old, adolescents start to believe in the need to stand out, in order to attract attention. On their search for this self- identity, adolescents start to experiment with and inspect their boundaries (Nelis & Van Sark, 2014). With the context of the developmental stage, it seems that additional explanations for the decline of Twitter are likely. Therefore, even though several of the participants agreed on this point with Mirthe, the decline of Twitter might not only be age-related. It is also possible that Twitter as a medium is not new and exiting anymore and new possibilities arise.
The male respondents also mentioned the good-for-nothing information as an important logic behind the decline of Twitter.
Erik: I have never used Twitter. To know that somebody else was using the men’s room or something, was pointless to me.
Not all participants have moved away from Twitter, but they do agree that the use of the social network site has changed. As Jojet says:
Jojet: I still like to read the tweets of artists, but that is all. The information posted on Twitter is not really relevant. I don’t find it interesting and it does not really matter. I think Facebook is more interesting.
Jojet speaks about a shift in the way Twitter is used. When they were younger, they would post information on Twitter. But now they only use it to follow celebrities and artists. This indicates that these adolescents are shifting from a proactive use to a more passive use of Twitter and portray less self- disclosing behavior.
The digital environment (purpose of social media use). Connected to the previous dimension is the purpose of the use of these social media. Strikingly, most of the respondents
used social networking sites, such as Facebook, as a way to communicate with their friends and look at other people´s activities and pictures to kill time. But in addition, it has now become a way to reach out to friends or connect to others.
Aletta: if you don’t have friends, you can make friends through Instagram. That’s the way it goes, right?
This quote illustrates how this act is perceived as common by Aletta. The female participant acts nonchalant about adding strangers to Instagram to make friends, and is looking for confirmation from the other participants. So while Jojet has indicated that the behavior on Twitter shifts from a more proactive use to a more passive use, according to Aletta the opposite is true for Instagram. This shows that Instagram is gaining popularity as a social network site. This might be explained by the characteristics of the SNS’s; text- based versus photo- based.
Girl participants also indicated that they used social networking sites, because they were afraid that they would miss out on something, and because it has become a standard in their peer group.
Jojet: It is an easy way to reach out to friends, it has become the standard, everybody does it so why wouldn’t I.
Jojet: Young people are afraid to miss out on something and not being able to join in the conversation.
This quote illustrates the importance of the peer group during adolescence. Social networking sites are used as a communication medium and as a way to keep up with events. These
technologies are central to most of the young people’s lives. This is not always perceived as a positive effect:
Sophie: The 'meangirls- effect' (she said this and that) is transferred to social media for the most part, where you can find a whole spectrum of people who are presenting themselves in a
particular way. And everybody then has a certain opinion about this person.
This participant explains that she believes that social media have a big influence on our daily lives at a superficial and unrealistic level. In addition, this participant brings up the social media use of her high school:
Sophie: I feel cornered by the school, because they kind of promote the use of social media. Instead of teaching and providing information, they keep an eye on you.
Jojet: We have this Facebook- page of the school and once we had a field day, pictures appeared on this page without permission.
Maaike: Even picture of girls in bikini’s when we went swimming during this day.
Sophie: And another girl was pictured on a really big poster in the library, without her permission. School just is a risk factor, instead of educating us in online behavior.
These quotes illustrate a discussion of the behavior of the school online. The female participants are not at ease with the way the school puts information about and places pictures of them online.
Risks (Behavior towards online risks). Overall the participants did not understand the risk of adding strangers on Instagram or Facebook, but acknowledged that it is a common
practice in their age group. Even though all participants agreed that it is risky to add strangers on social media or did not understand the rationale behind this, some of the participants admitted that they had added a stranger in the past. The reason behind accepting these strangers usually is that they have seen them at a party once. They indicate that they believe it is harmless if you have some common sense and keep a distance. Furthermore, the boy participants believed that there is a difference between boys and girls, because girls are more vulnerable.
Alex: Some girls or a friendly guy I played soccer with once, you know.[…] For example, when you can find a lot of personal information on their profile and I see that he plays soccer as well.
With girls it is more about if they live in your neighborhood and I vaguely remember them from a party, then I will add them.
Robin: I do believe that girls are more likely to encounter risks online. There simply are [refers back to another participant] perverts on the Internet and they will not seek out boys. Not that everybody is a pervert of course, but I believe they will seek out girls before boys. But on the other hand, I don’t think that it is a really big risk, if you handle it properly you can easily avoid the risks.
One participant had a different reason for adding strangers to social networks. He explains that he plays an online game (League of Legends) against other online players. When he frequently encounters the same player, he sometimes adds this player on Skype. In this quote he talks about meeting these other players in the real world:
Thijs: I personally have invited people that I knew from Skype, but not in my home. Of course it is not completely the same thing [as randomly adding people on Facebook], but if you meet at a bus station for example, it does not consume a lot of my time and if he does not show up, it does not matter to me.
Eye- catching is the paradox between the perceptions of online risks and the actual behavior online. They have been engaging in actions that they first associate with risky behavior. These particular adolescents explain why they believe their actions are not risky, mostly because they had previous contact with this person. This is possibly explained by the process of building online relationships, as Sung and Mayer (2012) describe in the context of online classrooms. In order to establish a social relationship, sharing personal information is required. The
effectiveness of this process can be seen in the example of Thijs. He illustrates how a social relationship is established with a stranger based on common interests; the game League of Legends. It might also be the result of a cognitive dissonance process, in which the participant agreed that it was perceived as a risky behavior to add strangers on SNS, but this would not be risky in his case since he would have legitimate reasons to do so.
In addition, the privacy settings were discussed. Most of the participants limited the visibility to their account, but not all of them were aware of their own settings.
Geert: I think my account is only visible for friends, or ‘friends of friends’. Rens: I have no clue
Erik: I think others can see my posts, but not my personal information. Aletta: I think I have limited the visibility of my pictures.
Marjolein: I think my privacy setting is ‘friends of friends’. Mirthe: I am not sure, I thought I have the same setting.
Here the paradox between their perceived awareness and actual behavior becomes apparent. Most of the participants believe they are quite aware of online risks, but when you directly ask them about the privacy settings of their online accounts, a lot of them are unsure about this.
Attitudes during focus groups. As mentioned before, many quantitative studies have been conducted on the online behavior of young people. However, this approach alone does not provide in- depth understandings of the complexity of their behavior. Therefore, the qualitative approach of focus groups was used in order to identify this complexity and the dynamics of group interactions. The participants displayed a range of attitudes during the focus groups, which indicated that the participants were uncomfortable to discuss the topic in a group, for instance by making fun of others. There were also some differences between the male and female
participants. for example ‘showing off’ occurred much more in the male focus groups. A vivid example from a boy in the VWO focus group:
Patrick: I see boys growing taller and smarter than their parents. So the parents don’t need to preach to us all the time, because sometimes we [the younger generation] know better.
This again shows signs of 'rebellion’, as did Chris when discussing the influences on awareness of risk behavior. The only quote in the female focus groups was made by Amber when the discussion focused on talking to strangers on the Internet.
Amber: This afternoon I was talking to this guy online, no clue who he was. But the conversation did not go further than ‘how are you doing?’. Then the conversation just ends.
Most of the participants were hesitant to talk about online risk behavior at first. After the emphasis was made on the general attitudes and behavior instead of personal experiences, participants felt more at ease. Especially after the association game, when the participants had a clearer image of the definition of online risks. When the participants discussed their own behavior online, their attitude was often nonchalant and they assumed that this behavior was common. Sometimes the participants who acted nonchalant, were looking for recognition and confirmation of the other participants. This clearly shows that even though these adolescents think they are quite mature, the influence of peer pressure is still noticeable. While they are searching for their own identity, there is a thin line between confidence and insecurity especially with topics where they have not fully established their own boundaries (Nelis & Van Sark, 2014).
Several participants tried to focus the attention on someone else in the focus group by making fun of them. The following quote is from a discussion about who the participants are talking to online:
Alex: For example, [..] you meet a nice girl and it turns out she is a really sweet person and then you meet up with her.
Eva: so how does that work?
Alex: It depends, sometimes you see someone on Instagram and you start following her account and then you meet up.
This interaction shows the process of identity exploration in which most individuals conform themselves to a certain role, in this case Alex conforms to the role of the confident ladies man, and Geert acts like the joker in this group.
Eye-catching was the input of one female participant. Her attitude towards social
networking sites was totally opposite of that of the most participants. She was strongly against the use of social media such as Facebook.
Sophie: I deleted my Facebook account, because there was too much information on there. The only account I have is on Instagram, and I mostly use it to keep an eye on my little brother.
Sophie does acknowledge that she finds it fascinating what people post online, but she sees the unrealistic perceptions and expectations that are raised as the most important online risk.
Sophie: the feeling that you are participating in an alienating process. That you are presenting yourself in a private setting, when this is actually publicly visible. And you are constantly comparing yourself to others online. [...] I found it difficult to make a connection between the things my friends sent to me, and the things that I would send back to them and what the true meaning was of these messages.
This quote illustrates growing worries of the misperception of privacy online, albeit from only one participant.
In sum, adolescents mostly see social network sites as a way to communicate with their peer group, and as a way to kill time. These participants do not believe they are at risk online, because they grew up with the Internet and because of their common sense. Strikingly, when the adolescents are asked about their behavior online, it seems that they do engage in the risks they associate with this subject. Still, the participants thought of themselves as quite aware of the risks online. When the groups discussed privacy settings online, it was apparent that not all
the visibility of their account to 'only friends', but not all of the participants were aware of the settings of their own accounts. Another eye-catcher is the gap between the students who are positive about information gatherings at school, and the policy of schools on this subject. Not one participant was positive about the information provided at the schools at this point. As indicated by the police department Kennemerland, the police force has tried to collaborate with high schools in the area to organize an information gathering. But the schools are hesitant, since this is a private and sensitive matter. And most of all, they don’t want to give parents the
impression that they need to worry about their children’s behavior, just because it is a topic at the school. This shows there is a disagreement about who is responsible for the development of online awareness among adolescents. The police department and the adolescents themselves see a great opportunity for high schools, while the latter seems hesitant to engage in a discussion about the related sensitive topics.
Study 2 Content analysis Facebook- profiles Descriptive statistics
The sample of the content analysis consists out of 154 Facebook- profiles, with 89 profiles (57,8%) belonging to females and 65 profiles (42,2%) belonging to males.
The privacy settings of photos was set at partially private with 98 of the 154 participants. This means that with 98 participants the profile picture and cover picture album were visible, and more than one picture was uploaded into that album. Only 16 of the 154 participants have set their Facebook photos to private, which means that only one profile picture is visible and only one cover picture. This is a requirement for a Facebook profile. Table 1 shows the number of participants that chose a particular privacy settings for all parts of a Facebook profile.
Table 1 Descriptives privacy settings
Most of the participants (91,6%) have not listed their sex interest (male/female) in their profile. Among the participants that have listed their sex interest, it seems that there is a low difference between the sex interest in males (4,5%) and females (3,9%), but due to the very low frequency of this topic, this difference is negligible.
Of the 154 participants, there were 25 participants with pictures listed containing explicit nudity, which means that the participant is depicted in swimwear (females in bikini or bathing suit, males in swim shorts). With regards to sexual behavior depicted in pictures, two of the participants have very explicit sexual behavior that is visible in these pictures. Both of these profiles reflect teenagers who have pregnancy pictures on their profile. Furthermore, eight participants have pictures on their profiles that are coded as explicit; this means that these pictures show the participant explicitly kissing another person. And 17 of the participants have pictures that show mild sexual behavior, which means hugging another person or showing love interest in another person in the pictures.
The last specific focus of this content analysis was the reference to alcohol in pictures. Only one of the coded Facebook profiles depicted very explicit alcohol references in which the
participant was displayed in an active drinking act. In addition, seven participants showed explicit alcohol references, which refers to the participant holding a glass or can that clearly contains alcohol. This is coded as an indicator to drinking alcohol and has a high probability that
Privacy settings Private (no information
provided)
N=
Partially private (partially visible for non- friends)
N=
Public (completely visible for non- friends)
N=
General information 97 4 53
Photo’s 16 98 40
Interests 101 1 52
the participant engaged in an active drinking act, but since this is not clearly visible this display is coded separately from the active drinking act. Furthermore, 15 participants displayed mild alcohol references on their Facebook- profile, which means that visual references are made to alcohol, e.g. a bottle of alcohol on the table or other’s on the picture holding alcohol.
Nonparametric tests
While examining the distributions of the variables, normality test and histograms showed non- normal distributions. Consequently, nonparametric tests were executed in SPSS. These tests provide measures for non-normal distributed variables and small samples. Even though the majority of tests were executed with the total sample (N= 154), several tests were executed with only a part of the total sample depending from the variables that were used (e.g. ' Nudity in pictures’ is tested with 54 participants). These nonparametric tests are applicable to smaller samples, but large differences are needed in the data in order to obtain a significant effect and assume an difference in the population.
Hypothesis 1- Privacy settings
The first hypothesis states that gender has an influence on privacy protection behavior than men, meaning that female participants have more private profiles than men. This hypothesis is tested for the different parts of Facebook- profiles (photo’s, interest, and timeline). The privacy settings of the general information are not included in this test, because this variable includes different values than that of the other variables (the distinction between private and no
information listed by the profile owner is uncertain). We expect that gender is related to the total of privacy settings. Since the distribution of these variables is not normal or linear, we use a nonparametric test.
To test the relation between the total score on privacy settings and gender, a new variable is created which is named 'TotalPrivacy'. A factor analysis shows that all three of the questions
about privacy that we measure with this hypothesis (general information excluded), load on one component. These variables are rescaled based on an acceptable Cronbach’s Alpha of .67 with a mean of 1.99 (SD= .55). The Mann-Whitney U-test was conducted to look at the differences between two groups (males/females) in total privacy settings. The distribution of total privacy settings among males and females is non- normal, but the shapes are similar (Appendix E) The results indicate that privacy settings did not differ significantly between females and males (U= 2774, z=-.46, p= .65). Females had an average rank of 78.83, while males had a slightly lower average rank of 75.68. These results reject the hypothesis that there is a difference in privacy settings between males and females. Therefore, we cannot conclude from these results that females show greater privacy protection behavior than men.
Hypothesis 2- Information disclosure in general
The second hypothesis predicts that gender is related to the total score on self- disclosure. This variable is created by summing the scores on education level, residency, cell phone number, e-mail address, complete name, name of school, places where participant has been, events, favorites, added ' about me' description, is family listed?, and current relationship status. These variables where all measured on a known/unknown- or yes/no- answering scale. Gender, profile cover picture and sex interest in..? were measured on different answering scale and therefore excluded from this analysis. The new variable is named ' TotalSelfdisclosure'.
The Mann-Whitney U-test was conducted to look at the differences between two groups (males/females) in total self- disclosure. The distribution of total self- disclosure among males and females is non- normal, but the shapes are similar (Appendix E) The results indicated that self- disclosure did not differ significantly between females and males (U= 2776, z=-.43, p= .66). Females had an average rank of 76.19, while males had a slightly higher average rank of 79.29.
These results reject the hypothesis that there is a difference in amount of self- disclosure between males and females.
In addition, a Kruskal- Wallis H-test was conducted between Total privacy settings and Total self- disclosure to test the hypothesis that the amount of information given on a profile differs between the different privacy settings (public, partially private, and private). The results showed there is a statistically significant difference in self- disclosure between public, partially private, and private profiles, X2 (2)= 73.07, p <.001, with a mean rank self- disclosure of 51.96 for private profiles, 103.61 for partially private profiles, and 131.41 for public profiles. These findings are in line with the expectations, since there is usually less information found on a private profile than on a public profile.
Hypothesis 3- Current relationship status
This hypothesis predicted that gender is related to the listing of current relationship status. This variable was tested with a single item about the listing of the relationship status in the general information- section.
The Mann-Whitney U-test was conducted to look at the differences between two groups (males/females) in current relationship status. The distribution of the listing of the relationship status among males and females is not normal, but the shapes are similar (Appendix E) The results indicated that the listing of the current relationship status did not differ significantly between females and males (U= 2837.5, z=-.33, p= .74). Females had an average rank of 76.88, while males had a slightly higher average rank of 78.35. These results reject the hypothesis that there is a difference in the listing of the current relationship status on a profile between males and females.
To test the relation between the level of privacy and the listing of current relationship status a Kruskal- Wallis H-test was conducted. The variable Total Privacy settings was used in this
