Does gender matter in the virtual world&quest ... - IngentaConnect

3 downloads 0 Views 117KB Size Report
Jun 17, 2013 - aDepartment of Criminal Justice, Shippensburg University, 1871 Old Main ... of Criminal Justice, Weber State University, 1206 University Circle, ...
Original Article

Does gender matter in the virtual world? Examining the effect of gender on the link between online social network activity, security and interpersonal victimization Billy Hensona,*, Bradford W. Reynsb and Bonnie S. Fisherc a Department of Criminal Justice, Shippensburg University, 1871 Old Main Drive, Shippensburg, PA 17257, USA. E-mail: [email protected] b Department of Criminal Justice, Weber State University, 1206 University Circle, Ogden, UT 84408-1206, USA. E-mail: [email protected] c School of Criminal Justice, University of Cincinnati, 665R Dyer Hall, ML 310289, Cincinnati, OH 45221-0389, USA. E-mail: [email protected]

*Corresponding author.

Abstract The growth in the number and scope of private information breaches of social networking (SN) websites subscribers’ has prompted researchers to focus on the utilization of privacy and security measures among SN users. Previous research suggests that users who engaged in risky online behaviors, such as adding strangers as friends, were more likely to be victimized online. The current study examines the role gender plays in the relationship between SN security and online interpersonal victimization (OIPV). Utilizing a probability sample of college students from a large Midwestern university, gender differences in online SN activity and security were found. Some factors (number of SN updates and adding strangers as friends) had a significantly different impact on OIPV among males and females. Security Journal (2013) 26, 315–330. doi:10.1057/sj.2013.21; published online 17 June 2013 Keywords: gender; social networking; online security; online interpersonal victimization

Introduction Online social networking (SN) programs (for example, Facebook, Twitter, LinkedIn) have created platforms for individuals all over the world to connect and share information, ideas and experiences. In a world full of seemingly never-ending technological advances, the popularity and growth of online SN websites is unprecedented. Owing to their widespread use and too frequent abuse by individuals, these websites have become key targets for criminology and victimology researchers. One central topic ripe for inquiry among academics and policymakers is the relationship between online SN use and online interpersonal victimization (OIPV).

© 2013 Macmillan Publishers Ltd. 0955-1662 Security Journal www.palgrave-journals.com/sj/

Vol. 26, 4, 315–330

Henson et al

OIPV is a form of victimization in which individuals are harassed, sexually solicited and/ or tormented by offenders online. It often includes various online or digital activities, such as repeated E-mails, instant messages, text messages, messages posted on SN sites, images or videos. Over the past decade, researchers’ interests have centered on specific types of OIPV, including cyberstalking, cyberbullying and online harassment. However, overlooked in their focus has been a gendered approach to OIPV. Research suggests that there is a strong gendered effect in explaining OIPV (Henson et al, 2011). Simply put, females may be more at risk than males when it comes to being victimized on the Internet. From an opportunity perspective, there may be gender-specific pathways to OIPV, a hypothesis that has been supported for offline forms of victimization (for example, Wilcox et al, 2009), crime and delinquency (for example, Miller, 1998; Heimer and De Coster, 1999; Belknap and Holsinger, 2006; Daigle et al, 2007), and fear of crime (for example, Schafer et al, 2006; May et al, 2010). The premise of this perspective is that the opportunity structures that contribute to OIPV among females may be different from those influencing victimization among males. As yet, this possibility, although theoretically sound, has not been empirically explored by more than a few studies. The current study attempts to bring together these perspectives on victimization to address the possibility of gendered opportunities for OIPV among college students.

Online SNs Online SN sites are online platforms that allow users to post personal information and pictures, link accounts and communicate with others, and form virtual communities on the Internet. Depending on the type of SN platform, users have a wide array of options, including uploading video or audio files, creating specialized groups and contacting potential employers. Since their initial introduction in the late 1990s, SN sites have grown to become part of the everyday lives of millions of people worldwide.1 According to a Nielsen Company’s (2009) study, nearly two-thirds of the global online population use SN and blog sites, making up almost 10 per cent of all time spent on the Internet. That study also reported that online communities (that is, SN sites) have become a more popular form of online communication than E-mail. Use of online SN sites is especially widespread among college students. For example, according to Facebook’s Social Ad Platform (2011), in January of 2011, collegeage users (18–24 years old) made up the largest group, accounting for over 30 per cent of all users. Further, almost 12 million users specifically noted they were enrolled in college. Given the nature and popularity of online SN sites, criminologists and victimologists have begun to recognize them as potential arenas for offending and victimization. For example, Debatin et al (2009) examined negative experiences and privacy issues of students who use Facebook at a large, Midwestern university. They found that almost 18 per cent of respondents reported negative experiences on their Facebook account over the course of their lifetime, including unwanted sexual advances, harassment and stalking. In a similar study, Henson et al (2011) examined OIPV among students at a different large, Midwestern university who used a wide array of SN sites such as Facebook, Myspace and Twitter. They reported that almost 42 per cent of respondents in their sample had experienced some form of OIPV over the course of their lifetime, including online harassment, unwanted sexual advances and/or threats of violence. 316

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

Online privacy protection One aspect of online SN that has received considerable attention among both researchers and the media is the issue of privacy protection. Although every SN site has some form of privacy protection, those settings can easily become ineffective because of the user’s actions. To illustrate, in one of the first studies to examine privacy issues on SN sites, Jones and Soltren (2005) reported that privacy among Facebook users is often undermined by the amount of information they disclose on their profiles. In their study on the behavior of SN site users, Acquisti and Gross (2006) found that although privacy was repeatedly mentioned as one of the biggest concerns, many respondents still posted personal information on their SN account. For example, 24 per cent of respondents reported they posted their home address, whereas 39 per cent posted their cell phone number. Similarly, Henson et al (2011) examined the relationship between SN security and OIPV. They found that while about 82 per cent of respondents set their SN accounts to private (meaning only individuals with linked accounts could view it), they still engaged in other risky online behaviors, such as adding strangers as friends (72 per cent of respondents), which made them more likely to be victimized online. As indicated by these and other studies, the desire for privacy and security among SN users may be undermined by the users’ own actions.

OIPV OIPV takes many forms, and while there is no standardized definition among academic researchers of what it entails, much like its spatial form, there are component behaviors that characterize this form of victimization. OIPV typically involves some form of contact on the Internet (for example, E-mail or instant messaging) or through other electronic means (that is, text messaging). This contact can be a single incident or repeated incidents, and usually involves communication from the offender to the victim in the form of E-mails, text messages, instant messages or other similar methods (for example, Facebook messages, message board posts). The types of behaviors experienced by victims may include harassment, sexual solicitations, intimidation or threats of violence. As noted by Henson et al (forthcoming), although OIPV may lack the physical component often associated with its spatial counterpart, it may still produce fear and emotional stress. College students have been identified as an at-risk group for many types of online victimization, such as computer virus infections, hacking, cyberstalking, cyberbullying, harassment and unwanted sexual solicitation (for example, Choi, 2008; Holt and Bossler, 2009, 2010; Kraft and Wang, 2010; Marcum et al, 2010a; Henson et al, 2011). While certain types of personal victimization among college students appear to have declined in the last few years2 (Baum and Klaus, 2005; Katel, 2011), victimologists continue to recognize the importance of studying this population as well as studying the college campus as a context for victimization (Fisher et al, 2010; Sloan and Fisher, 2011). In this light, recent work has examined the extent of OIPV among college students (for example, Spitzberg and Hoobler, 2002; Kraft and Wang, 2010; Reyns et al, 2011). However, the rather limited number of studies that have been published highlights the need for more research in this growing area. Despite the methodological limitations of some of these studies (for example, small sample sizes, convenience samples), they nevertheless offer a © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

317

Henson et al

foundation upon which to build future studies examining the extent and nature of college student OIPV.

OIPV among college students The growing body of OIPV research focused on college students has mostly explored cyberstalking, harassment and unwanted sexual exposure. This research indicates that it is experienced by a significant portion of students, and that certain online behaviors related to SN use can increase victimization risks (Henson et al, 2011; Reyns et al, 2011). In one of the first studies to examine cyberstalking victimization, Spitzberg and Hoobler (2002) analyzed data from a sample of 235 undergraduate communication students at a Southwestern public university. The authors reported that 1–31 per cent of the sample experienced cyberstalking at some point in their lives, depending upon the type of pursuit behavior under consideration.3 In a similar study, Kraft and Wang (2010) surveyed college students at a public liberal arts college, and reported that among the 471 respondents, 9 per cent had experienced cyberstalking and 10 per cent had experienced cyberbullying while in college. Reyns et al (2012) also found high victimization rates among their sample of college students – 41 per cent of their sample had experienced some form of cyberstalking during their lifetime, such as online harassment or online threats (see also Reyns et al, 2011). This study also revealed that activities such as owning multiple SN accounts, updating them often, using instant messaging programs, friending strangers, using profile trackers and engaging in online forms of deviance (for example, hacking, piracy, sexting) all significantly increased students’ likelihood of cyberstalking victimization. The online harassment literature similarly suggests that college students are at risk for victimization. For example, in an early study of online harassment, Finn (2004) examined victimization among a sample of students at the University of New Hampshire. He asked respondents if they had received any threatening, insulting or harassing messages online. On the basis of his findings, Finn estimated that approximately 10 per cent of undergraduate students at the university had been a victim of online harassment while in college. A replication of Finn’s study by Lindsay and Krysik (2012) reported a 43 per cent rate of victimization while in college, and that time spent on online SN sites increased victimization risk. Further, Lindsay and Krysik also reported a significant relationship between sexting – sending nude or semi-nude pictures of oneself via text message – and online harassment. Interestingly, Reyns et al (2013) reported similar findings with respect to cyberstalking victimization. They found that respondents who participated in sexting not only were more likely to experience online victimization but were also more likely to experience multiple types of victimization behaviors (for example, harassment and unwanted sexual advances). Another influential study of the online harassment of college students was conducted by Holt and Bossler (2009) (see also Holt and Bossler, 2010). Study participants were asked how many times in the last 12 months they had been harassed in a chatroom, Internet relay chat or instant message chat. It was found that almost 19 per cent of students surveyed reported they had been harassed online. While this study asked only about use of the SN Myspace and did not uncover any significant relationships with victimization, other online behaviors such as visiting chatrooms and engaging in hacking were related to online harassment. Cybercrime researchers also have been investigating the nature of sexual forms of OIPV. For example, Marcum et al (2010a) investigated forms of online sexual victimization 318

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

among college students, including receiving sexually explicit materials online and online sexual solicitations. The authors reported that 13 per cent of students received unwanted sexual materials online and 6 per cent were sexually solicited while online. As to online SN usage, Marcum et al (2010a) reported that the use of Facebook was significantly but negatively related to victimization, meaning those who used Facebook were less likely to receive sexually explicit materials, but posting personal information on SNs increased risks for this type of victimization. Neither of these behaviors, however, was significantly related to online sexual solicitation. Reyns et al (2012) also reported that 13 per cent of their sample of college students were victims of unwanted sexual advances while online, but in a subsequent study, the authors found that SN activity was not related to online sexual victimization (Reyns et al, 2011). Collectively, these and other studies, have revealed that a somewhat substantial percentage of college students experience OIPV.

Gendered Opportunities for OIPV Some of the previously reviewed research studies of OIPV among college students tested the proposition that pathways to victimization may vary with gender. Although the possibility of unique gendered pathways to victimization is intriguing, it is not unique to the cybervictimization literature (for example, Hindelang et al, 1978; Wilcox et al, 2009; Henson et al, 2010; Popp and Peguero, 2011). One of the central propositions of Hindelang et al’s (1978) lifestyle-exposure theory is that gender influences lifestyles, which in turn affect victimization risk. According to the theory, exposure to risk for personal victimization varies by lifestyle. Moreover, demographic characteristics, including age, sex, race, income, marital status, education, and occupation are the antecedents to lifestyle. Hindelang et al (1978) argue that sex role socialization is the primary factor behind gender differences in lifestyles, and hence also differences in personal victimization: ‘Because of sex role differentiation, sex is related to daily activities such as where time is spent, the number of interpersonal contacts, and the likelihood of encountering strangers’ (p. 248). In an online realm, the number of interpersonal contacts and the likelihood of encountering strangers are potentially limitless depending upon one’s online routine activities, but nevertheless may vary by gender. Building on this premise, a small body of cybervictimization research suggests that influences on victimization may be conditioned by gender (for example, Holt and Bossler, 2009, 2010; Marcum et al, 2010a). For instance, Holt and Bossler (2009) examined the effects of online routines and guardianship measures of a sample of college students on their likelihood of malware infection. The authors reported that males who spent their time online programming or downloading (pirating) media, were at increased risk for malware infection, whereas females were more likely to be victims of malware infection if they visited chat rooms (Holt and Bossler, 2009). In another study by Holt and Bossler (2009), the predictors of online harassment were examined separately for male and female college students. The authors reported that for both groups, using chat rooms was a precursor to online harassment, but among females, friends’ deviance (for example, hacking, viewing pornography, pirating media) also increased their likelihood of malware infection. In addition, Marcum et al (2010b) expanded their previously discussed analysis to examine exposure to sexually explicit materials, nonsexual harassment and sexual solicitation among college students based on gender. The authors reported different influences on victimization for males and © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

319

Henson et al

females.4 For instance, E-mailing increased the likelihood of sexual solicitation among males, while instant messaging increased females’ risk. The groundwork for an exploration into the relationships between online SN activity, privacy and OIPV has been laid, and the previously reviewed studies support the hypothesis that online behaviors may be different for males and females. Yet, thus far researchers have not addressed the role that gender plays as it pertains to online SN usage and the genesis of OIPV. The purpose of the present study, therefore, is to bridge these three areas of research by examining the relationship between online SN activity, security and OIPV among males and females.

The Present Study In a previous study examining the link between the use of online SNs, SN security and OIPV, we reported that gender was one of the key variables of interest (Henson et al, 2011). The current study will extend our previous work and focus on potential gender differences in SN behaviors and OIPV. In doing so, we provide some answers to three primary research questions: 1. Does online SN activity differ between males and females? 2. Does the use of online security behaviors differ between males and females? 3. Do online SN activity and security measures differentially affect the likelihood of OIPV for males and females?

Methods Data collection Data for this study were collected as part of a larger study examining online victimization among college students. All data were obtained through the use of a self-report Web-based survey. To improve the format and structure of the survey, design elements were adapted from both Dillman’s Tailored Design Method and previous online surveys (Dillman, 2007; Dillman et al, 2008). Before administering the survey, a pilot test was conducted with students in a Master’s level criminal justice research methods class. Students’ comments were used to address technical issues and improve the likelihood of respondent comprehension and completion or survey questions. Once a sample population of potential participants was obtained (see sample selection below), the university registrar’s staff, with consent from the university Internal Review Board, E-mailed each student an invitation asking them to participate in a study about their experiences with online victimization. The E-mail described the study’s protocol and the methods of protections for participant confidentiality. It also instructed respondents to click on an embedded link if they were interested in participating in the study. The first page of the survey was an information sheet, which concluded with a statement alerting respondents that by continuing with the survey they were giving consent for their information to be used for research purposes. Two additional waves of follow-up E-mails were sent 2 and 4 weeks after the initial invitation. In an effort to 320

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

prevent the data from being influenced by individuals completing the survey multiple times, the survey was designed so individuals could only complete it once. Sample selection During the spring quarter of 2009, a simple random sample of 10 000 undergraduate college students at a large urban university in the Midwest was obtained with the aid of the university’s registrar’s office.5 To be included in the sample, students must have been: undergraduate, attending full-time and between the ages of 18 and 24. According to the Digest of Higher Education Statistics, this population is representative of the typical college student (US Department of Education, 2012). Owing to issues with some E-mail addresses, 9926 invitations to participate in the online survey were sent via E-mail to students. Of those who were sent an invitation, 1310 students participated in the survey, resulting in a response rate of 13.1 per cent. As the goal of the present study is to examine respondents’ experiences with SN sites, we chose to analyze only those respondents that reported having a SN account. In addition, two respondents were removed because they did not signfy their gender. As a result, the analytical sample for the current analysis includes 906 respondents.6 Although the response rate obtained may be low by traditional standards, it is not unusual given the mode of survey administration and population under study (Couper, 2000). Furthermore, Dillman et al (2009) conducted a study to assess response rates of several different methods of survey administration, concluding that whereas mail surveys had response rates as high as 75 per cent, Web-administered surveys had the lowest response rates at 12.7 per cent. Other studies examining similar populations and utilizing similar survey methods have also reported response rates comparable to that in the current study (for example, Hilinski, 2009; Nobles et al, 2009; Patton et al, 2010). Dependent measures As discussed previously, OIPV has been operationalized using a number of different types of behaviors, including harassment, unwanted sexual advances, intimidation and threats of violence. To obtain a comprehensive measure of OIPV that has appropriate content validity, respondents were questioned about their online experiences with unwanted contact, harassment, unwanted sexual advances, and/or threats of violence or physical harm. For each type of behavior, respondents were asked if they had ever experienced it on multiple occasions (two or more times) on their SN sites, after telling the perpetrator to stop. Each of the four types of behaviors was coded as ‘yes’ or ‘no’. The number of ‘yes’ responses were then summed across the four types of behaviors and recoded as a single dichotomous measure of OIPV, indicating if the respondents had ever experienced any of the behaviors (Table 1). Independent measures SN activity The current analysis includes six SN information and activity measures: (i) number of online SN accounts; (ii) time spent on SN; (iii) number of SN updates; (iv) number of SN friends; © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

321

322 Henson et al

© 2013 Macmillan Publishers Ltd. 0955-1662

Table 1: Variables, scales, and descriptive statistics Variables

Scale/Coding

Security Journal

Dependent variable Experienced OIPV

0 = No, 1 = Yes

Independent variables SN activity Number of SN accounts Number of time on SN Number of updates Number of friends on SN account(s) Number of photos on SN account(s) Flirting on SN account(s)

Total number of accounts opened Number of hours per day Number of times updated per week Total number of friends Total number of photos 0 = No, 1 = Yes

Vol. 26, 4, 315–330

Security actions Set account to private Added strangers as friends Used friend service Used profile tracker

0 = No, 1 = Yes 0 = No, 1 = Yes 0 = No, 1 = Yes 0 = No, 1 = Yes

Controls Race Age Sexual orientation Relationship status

0 = White, 1 = Non-White Age in years 0 = Heterosexual, 1 = Non-heterosexual 0 = Single, 1 = Non-single

Mean

Standard deviation

Minimum

Maximum

N

0.417

0.493

0

1

906

2.600 1.580 2.410 513.500 350.370 0.500

1.700 1.722 3.852 524.734 466.764 0.500

1 1 0 7 0 0

15 16 25 5000 5000 1

897 904 902 885 878 906

0.820 0.720 0.030 0.120

0.387 0.450 0.182 0.327

0 0 0 0

1 1 1 1

903 901 905 904

0.123 20.166 0.058 0.577

0.329 1.317 0.233 0.494

0 18 0 0

1 24 1 1

900 906 901 903

Does gender matter in the virtual world?

(v) number of photos posted online; and (vi) flirting on SN. The number of SN accounts is the total number of active SN accounts the respondents had with any SN provider. Time on SN is the average number of hours per day a respondent spent viewing, updating and communicating with others on a SN account. The number of updates represents the number of times per week, on average, that users updated the information on their SN account. The number of friends is the total number of individuals linked to a respondent’s account as ‘friends’. Similarly, the number of photos is the total number of photos or images posted to a respondent’s SN account. Finally, the flirting on SN measure was creating by summing the dichotomous responses (yes = 1, no = 0) to two survey questions: (i) Have you ever flirted with a friend/acquaintance on your SN account? and (ii) Have you ever flirted with someone you did not know on your SN account? The resulting variable was recoded as ‘yes’ or ‘no’, with ‘yes’ meaning the respondent had flirted with someone – either a friend or stranger – and ‘no’ meaning the respondent had never flirted with someone on their SN account. The coding schemes and descriptive statistics for each SN activity variable are displayed in Table 1. SN security The current analysis also includes measures of SN security behaviors, including: (i) set account to private; (ii) added strangers as friends; (iii) used friend service; and (iv) used profile tracker. Set account to private indicates whether the respondent had his/her SN account set to private access. Added strangers as friends specifies whether the respondent added individuals they did not previously know as friends. Used a friend service indicates whether a respondent joined an online service that assists with getting large numbers of new friends. With such a service, individuals enter their SN ID or name into an online program, which sends friend requests to a large number of other SN users, with the goal of quickly increasing one’s number of friends. Finally, used profile tracker signifies whether a respondent used a program designed to track and report the names of individuals who visited his/her SN profile. Each of these four measures is dichotomous, coded as ‘yes’ or ‘no’. The coding schemes and descriptive statistics for each security behavior variable are displayed in Table 1. Control measures Finally, several control measures also are utilized in the current analysis, including measures of race, age, sexual orientation and relationship status. Race (White or non-White), sexual orientation (heterosexual or non-heterosexual) and relationship status (single or non-single) are each dichotomous variables, while age is a continuous variable (18–24 years). As the main goal of this study is to examine gender differences, gender is not being used as a control variable. Table 1 displays the coding schemes and descriptive statistics for each control variable.

Results As a first step in examining potential gender differences in online SN activity, security and OIPV, an independent t-test was performed. Displayed in Table 2, several variables were found to be significantly different for male and female respondents. In terms of the SN © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

323

Henson et al

Table 2: Comparison of online SN activity, security and interpersonal victimization by gender Males (N = 347)

SN activity Number of SNs Number of SN updates Time on SNs Number of friends on SNs Number of pictures on SN Flirting on SN SN security Set account to private Use friend service Used profile tracker Add strangers as friends Experienced OIPV

Females (N = 559)

Mean

Standard deviation

Mean

Standard deviation

2.487 2.488 1.351 490.787 208.935 0.568**

1.953 3.951 1.454 543.761 286.321 0.496

2.673 2.351 1.710** 528.199 437.475** 0.454

1.525 3.776 1.854 513.269 532.061 0.498

0.442 0.183 0.254 0.428 0.470

0.869** 0.034 0.154** 0.694 0.472**

0.338 0.182 0.361 0.461 0.500

0.734 0.035 0.069 0.760* 0.329

t valueðdf Þ x2 valueðdf Þ†

−1.589 (895) 0.522 (902) −3.065 (900) −1.032 (883) −7.273 (876) 11.002 (1)

25.983 (1) 0.001 (1) 14.463 (1) 4.684 (1) 18.193 (1)

*P ⩽ 0.05; **P ⩽ 0.01. † A χ2 analysis was performed to determine if a statistically significant difference existed between males and females for flirting on SN, set account to private, used friend service, use profile tracker, add strangers as friends and experienced OIPV, because the variables are categorical rather than continuous.

activity measures, time on SN, number of pictures on SN and flirting on SN were all significantly different. For the time and pictures measures, the means were higher for female respondents, while the mean for the flirting measure was higher for male respondents. With regard to the SN security measures, set account to private, use of profile tracker and added strangers as friends were all significantly different between males and females. For the private and tracker measures, the means were higher for female respondents, while the mean for the added strangers measure was higher for male respondents. Finally, there was also a significant difference between male and female respondents for the OIPV measure, with female respondents having a higher mean level of victimization. To provide more detailed information about the potential gender differences in the relationship between SN activity, SN security and OIPV, binary logistic regression models were estimated for male and female respondents separately. As illustrated in Table 3, there are several significant relationships for both the male and female respondents. For male respondents, the number of SNs and the number of pictures on SNs are associated with a statistically significant increase in the likelihood of OIPV. Further, male respondents who have added strangers as friends are almost six times more likely to experience OIPV. For female respondents, number of SNs, number of SN updates and time on SNs are associated with a statistically significant increase in the likelihood of OIPV. In addition, female respondents who reported flirting on SNs, using a profile tracker, and adding strangers as friends are two times more likely to experience OIPV. Given the number of significant measures in both the male and female regression models, a test of equality of regression coefficients was performed. The goal was to determine which, 324

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

Table 3: Binary logistic regression models for OIPV by gender Males

Females

Test of equality of regression coefficients

Coefficient Standard Exp(b) Coefficient Standard Exp(b) error error

z score

SN activity Number of SNs 0.181* Number of SN updates −0.011 Time on SNs 0.096 Number of friends on SNs 0.000 Number of pictures on SNs 0.001* Flirting on SN 0.521

0.074 0.035 0.094 0.000 0.000 0.280

1.198 0.989 1.101 1.000 1.001 1.683

0.180* 0.092** −0.145* 0.000 0.000 0.751**

0.078 0.034 0.600 0.000 0.000 0.200

1.197 1.096 0.865 1.000 1.000 2.120

0.009 −2.111* 0.397 0.000 0.000 −0.668

SN security Set account to private Used friend service Used profile tracker Add strangers as friends

0.111 −0.566 0.667 1.742**

0.304 0.695 0.487 0.440

1.117 0.568 1.949 5.707

0.330 0.615 0.724* 0.660**

0.284 0.588 0.292 0.212

1.391 1.849 2.062 1.936

−0.526 −1.297 −0.100 2.215*

Controls Race Age Sexual orientation Relationship status Constant

0.156 0.433 0.217 0.054 −6.591**

0.109 0.430 0.562 0.268 2.284

1.542 1.169 1.242 1.056 0.001

0.073 0.299 0.407 0.204 1.512

1.243 1.023 1.476 1.354 0.069

1.014 0.411 −0.248 −0.739 −1.429

0.023 0.218 0.389 0.303 −2.678

*P ⩽ 0.05; **P ⩽ 0.01.

if any, of the relationships between the SN activity, SN security and OIPV measures were significantly different for male and female respondents. The results can be seen in the final column of Table 3. The relationship between the number of SN updates and OIPV and the relationship between adding strangers as friends and OIPV were significantly different for male and female respondents. This finding indicates that there is a clear gender difference in the influence of each of the two measures on the probability of ever experiencing OIPV. The significance of these relationships is discussed in the next section.

Discussion and Conclusions The purposes of this study were threefold. First, it was necessary to determine if SN activity differed significantly between males and females. As discussed, statistically significant differences were uncovered between males and females for several types of online SN activity, such as the amount of time spent using online SNs, the number of photos posted on SN accounts and flirting on SN accounts. Second, we examined the extent to which online security behaviors varied between males and females. The results again indicated statistically significant differences related to account privacy settings and use of profile trackers. Interpreting these findings through the lens of opportunity theory, differential online SN © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

325

Henson et al

behaviors create unique opportunity structures for victimization, and should theoretically lead to different victimization risks for males and females. Having confirmed that males and females differ in their online SN activities and security behaviors, the third purpose of the present study was to assess the effects that these differences had on ever experiencing OIPV. Among both males and females, two online SN behaviors were identified as increasing students’ risks for victimization. Specifically, the number of SNs and adding strangers as friends to online SNs were related to an increased likelihood of ever experiencing OIPV. Again, opportunity theory and the lifestyle-routine activities perspective are useful in interpreting these results. According to lifestyle-routine activities theory (LRAT), increased exposure and proximity to motivated offenders increases victimization risk (Cohen et al, 1981). These two online SN behaviors – number of network accounts and adding strangers – can be considered dimensions of these two theoretical concepts, respectively (Reyns et al, 2011). Therefore, these findings are in accord with theoretical expectations. Moreover, although the current study should not be considered a full rigorous test of the LRAT perspective is still useful in interpreting the results of this analysis. Given the differential opportunity structures for victimization created by differences in online behaviors between males and females, opportunity theory suggests that males and females should have different pathways to victimization. In support of this idea, several factors were unique predictors of OIPV between the genders. To begin, the number of pictures posted on online SNs was related to victimization among males, but not females. Among males, increases in the number of photos posted translated to an increase in victimization risk, an effect that is once again explained as an element of exposure under LRAT. With respect to unique influences on victimization against females, the number of SN updates, the amount of time spent in these networks, flirting online, and use of profile trackers were all significantly related to OIPV, but in different ways. That is, the number of updates, flirting, and using profile trackers each increased victimization risk for females by as much as two times, while the amount of time spent within online SNs decreased these risks. In other words, the more time female students spent engaged in online SN resulted in a decrease in likelihood of victimization. These results each warrant a brief explanation. First, the number of updates could be conceived as a reflection of online exposure in which case the observed increases in victimization risk, although modest, are supported by theory. Second, flirting is a behavior that usually communicates to the recipient an interest in some type of relationship. Utilizing lifestyle-routine activities as a guide, this behavior could be indicative of target suitability, especially for the types of victimization included in our operational definition of OIPV (for example, sexual advances, harassment). Future research should continue to investigate this effect, as the current study may be the first one to do so. Third, use of a profile tracker, which may indicate self-guardianship, was significantly and positively related to OIPV. As discussed elsewhere (Henson et al, 2011), this finding is likely the consequence of a temporal ordering issue, wherein individuals utilized these programs in response to problems rather than prior to their victimization. Finally, the test of equality of regression coefficients revealed that among these gendered effects on victimization, there were significant differences between males and females related to the number of SN updates and adding strangers as friends. 326

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

Limitations and Future Research Like much research, the current study is not without its shortcomings, and the utility of the results are potentially impacted by three methodological limitations. Ideally, future research will not only replicate this study, but also improve upon these deficiencies. First, crosssectional data were utilized to answer our three primary research questions. Although crosssectional data are often the norm in victimization research and no publicly available longitudinal data have been collected that could have served our purpose, it is still the case that longitudinal data are preferred in order to establish time order and therefore causality. Thus, this should be a focus of future research. Second, as was discussed previously, the response rate for the survey was somewhat low. Although online surveys are notorious for this issue, higher response rates are preferable to minimize other methodological problems, such as response bias. On the other hand, online methods of survey administration were ideal for this study given the high prevalence of Internet usage among college students and the nature of the research questions. A challenge for the future, therefore, will be to elicit high response rates among this population using online survey methods. Third, the sample of college students that the analyses and conclusions were based on was drawn from a single university setting. To improve generalizability, replications of this research would benefit from examining representative samples of college students from multiple campuses.

Implications for Theory, Research and Prevention Limitations aside, results from the current study have implications for theory, research and prevention. The idea of a gendered opportunity theory has recently been gaining the attention of researchers studying victimization (for example, Wilcox et al, 2009; Henson et al, 2010). The growing body of research suggests that gender may be a trait by which opportunities are created and/or altered. However, the theoretical and analytic approach to studying gendered opportunities has varied across research studies. One vein of research suggests that gender structures lifestyles and routine activities – and therefore influences opportunities for victimization (Hindelang et al, 1978). In other words, gender influences lifestyle, lifestyle affects opportunity and opportunity facilitates victimization. Yet, another approach has been to consider how gender interacts with other risk factors for victimization. In this case, the theoretical argument is that lifestyles and routines will have differential effects on victimization based on whether the potential victim is male or female, regardless of whether the activities themselves are male-centric or female-centric. The current study, by examining differences in online activities between males and females and estimating the effects of those routines, has implications for each of these gendered perspectives. That is, based on our analysis it appears that males and females differ in their participation, time invested, protection and activities while online. We cannot, however, say whether these differences are because of gender. While plausible explanations of gender differences in lifestyles and routine activities have been offered (see Hindelang et al, 1978), no theoretical or empirical work to date has addressed why gender should influence online behaviors, lifestyles or routines. Further, while we report that certain online behaviors (for example, flirting, number of SNs) affect OIPV differently for males and females, we cannot determine whether this is because of gender or due to differences in nature of the online activities © 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

327

Henson et al

themselves. For example, does flirting on online SNs increase OIPV risk because offenders are more likely to target females (see Wilcox et al, 2009), or because of the types of networks, offenders or styles of online flirting that females employ in comparison to males? Such questions could not be addressed in this study. The current study, nonetheless, contributes to the growing body of work investigating gendered opportunity, and further suggests that these theoretical approaches apply equally to online forms of interpersonal victimization. Results of the current study also have implications for security and prevention of OIPV, especially within the domains of online SNs, which provide easy opportunities for motivated offenders to reach a nearly unlimited pool of suitable targets. In particular, our results indicate that allowing strangers access to SN profiles substantially increases victimization risk for both males and females. Opportunities for stranger-based OIPV could therefore be reduced if individuals were more cautious in their choice of online friends. However, this does not address the issue of OIPV perpetrated by those known to the victim. It is also noteworthy that certain SN behaviors, such as updating frequently or flirting were positively related to victimization. Yet, at some point, prevention strategies must be balanced with practicality, and although reducing these behaviors may also diminish victimization risks, there must also be ways to lower OIPV without unnecessarily constraining users’ behavior. Short of changing – and perhaps constraining – the legitimate behaviors of SN users, it may prove effective to utilize strategies involving bystander intervention, online community policing or online place managers to guard against OIPV.

Acknowledgements The authors would like to thank Kristin Swartz for helpful comments on an earlier version of this article.

Notes 1 In December 2012, Facebook, alone, had over 618 million active daily users worldwide (Facebook’s Social Ad Platform, 2011). 2 It is noteworthy that the sources indicating a decline in college student victimization (for example, the National Crime Victimization Survey) have not yet evaluated cybervictimization of college students. 3 For example, 1 per cent of respondents first met someone online (for example, through online dating websites) and then were subsequently stalked by that individual, whereas 31 per cent were sent exaggerated tokens of affection (poems, songs) online that implied obsessive pursuit. 4 Although the authors did report different influences on victimization for males and females, a true comparison is impossible because the authors’ male models include different variables than the female models. 5 Given the type of population and the inability to provide respondents with any potential incentives to participate in the study, we expected a low response rate. We chose to sample 10 000 individuals in order to obtain a sample large enough to perform appropriate statistical analyses. 6 An exclusion criterion was also used. Six cases were removed because they had missing data for more than half of the variables in the analysis.

References Acquisti, A. and Gross, R. (2006) Imagined communities: awareness, information sharing, and privacy on the Facebook. In Proceedings of the 6th International Workshop on Privacy Enhancing Technologies. 28–30 June, Lecture Notes in Comupter Science. Berlin-Heidelberg: Springer.

328

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

Does gender matter in the virtual world?

Baum, K. and Klaus, P. (2005) Violent Victimization of College Students, 1995-2002. Washington DC: U.S. Department of Justice. Belknap, J. and Holsinger, K. (2006) The gendered nature of risk factors for delinquency. Feminist Criminology 1(1): 48–71. Choi, K. (2008) Computer crime victimization and integrated theory: An empirical assessment. International Journal of Cyber Criminology 2(1): 308–333. Cohen, L.E., Kluegel, J.R. and Land, K.C. (1981) Social inequality and predatory criminal victimization: An exposition and test of a formal theory. American Sociological Review 46(5): 505–524. Couper, M.P. (2000) Web surveys: A review of issues and approaches. Public Opinion Quarterly 64(4): 464–494. Daigle, L.E., Cullen, F.T. and Wright, J.P. (2007) Gender differences in the predictors of juvenile delinquency: Assessing the generality-specificity debate. Youth Violence and Juvenile Justice 5(3): 254–286. Debatin, B., Lovejoy, J.P., Horn, A.K. and Hughes, B.N. (2009) Facebook and online privacy: Attitudes, behaviors, and unintended consequences. Journal of Computer Mediated Communication 15(1): 83–108. Dillman, D.A. (2007) Mail and Internet Surveys: The Tailored Design Method. Hoboken, NJ: Wiley. Dillman, D.A., Smyth, J.D. and Christian, L.M. (2008) Internet, Mail, and Mixed-Mode Surveys: The Tailored Design Method. Hoboken, NJ: Wiley. Dillman, D.A. et al (2009) Response rate and measurement differences in mixed-mode surveys using mail, telephone, interactive voice response (IVR) and the Internet. Social Science Research 38(1): 1–18. Facebook’s Social Ad Platform. (2011) Advertise on Facebook. https://www.facebook.com/advertising, accessed 22 January 2013. Finn, J. (2004) A survey of online harassment at a university campus. Journal of Interpersonal Violence 19(4): 468–483. Fisher, B.S., Daigle, L.E. and Cullen, F.T. (2010) Unsafe in The Ivory Tower: The Sexual Victimization of College Women. Thousand Oaks, CA: Sage. Heimer, K. and De Coster, S. (1999) The gendering of violent delinquency. Criminology 37(2): 277–318. Henson, B., Reyns, B.W. and Fisher, B.S. (2011) Security in the 21st century: Examining the link between online social network activity, privacy, and interpersonal victimization. Criminal Justice Review 36(3): 253–268. Henson, B., Reyns, B.W. and Fisher, B.S. (forthcoming) Cybercrime victimization. In: C.A. Cuevas and C.M. Rennison (eds.) The Wiley-Blackwell Handbook on the Psychology of Violence. Massachusetts: Wiley-Blackwell. Henson, B., Wilcox, P., Reyns, B.W. and Cullen, F.T. (2010) Gender, adolescent lifestyles, and violent victimization: Implications for routine activity theory. Victims and Offenders 5(4): 1–26. Hilinski, C.M. (2009) Fear of crime among college students: A test of the shadow of sexual assault hypothesis. American Journal of Criminal Justice 34(1): 84–102. Hindelang, M.J., Gottfredson, M.R. and Garofalo, J. (1978) Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, MA: Ballinger. Holt, T.J. and Bossler, A.M. (2009) Examining the applicability of lifestyle-routine activities theory for cybercrime victimization. Deviant Behavior 30(1): 1–25. Holt, T.J. and Bossler, A.M. (2010) The effect of self-control on victimization in the cyberworld. Journal of Criminal Justice 38(3): 227–236. Jones, H. and Soltren, J.H. (2005) Facebook: Threats to privacy, http://www.swiss.ai.mit.edu/6095/student-papers/ fall05-papers/facebook.pdf, accessed 30 March 2010. Katel, P. (2011) Crime on campus: Are colleges doing enough to keep students safe? CQ Researcher 21(5): 97–120. Kraft, E.M. and Wang, J. (2010) An exploratory study of the cyberbullying and cyberstalking experiences and factors related to victimization of students at a public liberal arts college. International Journal of Technoethics 1(4): 75–91. Lindsay, M. and Krysik, J. (2012) Online harassment among college students: A replication incorporating new Internet trends. Information, Communication & Society 15(5): 703–719. Marcum, C.D., Higgins, G.E. and Ricketts, M.L. (2010a) Potential factors of online victimization of youth: An examination of adolescent online behaviors utilizing routine activity theory. Deviant Behavior 31(5): 381–410. Marcum, C.D., Ricketts, M.L. and Higgins, G.E. (2010b) Assessing sex experiences of online victimization: An examination of adolescent online behaviors using routine activity theory. Criminal Justice Review 35(4): 412–437. May, D.C., Rader, N.E. and Goodrum, S. (2010) A gendered assessment of the ‘threat of victimization’: Examining gender differences in fear of crime, perceived risk, avoidance, and defensive behaviors. Criminal Justice Review June 35(2): 159–182. Miller, J. (1998) Up it up: Gender and the accomplishment of street robbery. Criminology 36(1): 37–66.

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330

329

Henson et al

Nielsen Company. (2009) Global faces and networked places: A Nielsen report on social networking’s new global footprint, http://blog.nielsen.com/nielsenwire/wp-content/uploads/2009/03/nielsen_globalfaces_mar09.pdf, accessed 30 March 2010. Nobles, M.R., Fox, K.A., Piquero, N. and Piquero, A.R. (2009) Career dimensions of stalking victimization and perpetration. Justice Quarterly 26(3): 476–503. Patton, C.L., Nobles, M.R. and Fox, K.A. (2010) Look who’s stalking: Obsessive pursuit and attachment theory. Journal of Criminal Justice 38(3): 282–290. Popp, A.M. and Peguero, A.A. (2011) Routine activities and victimization at school: The significance of gender. Journal of Interpersonal Violence 26(12): 2413–2436. Reyns, B.W., Henson, B. and Fisher, B.S. (2011) Being pursued online: Applying cyberlifestyle-routine activities theory to cyberstalking victimization. Criminal Justice and Behavior 38(11): 1149–1169. Reyns, B. W., Henson, B. and Fisher, B. S. (2012) Stalking in the twilight zone: Extent of cyberstalking victimization and offending among college students. Deviant Behavior 33(1): 1–25. Reyns, B.W., Burek, M.W., Henson, B. and Fisher, B.S. (2013) The unintended consequences of digital technology: Exploring the relationship between sexting and cybervictimization. Journal of Crime and Justice 36(1): 1–17. Schafer, J.A., Huebner, B.M. and Bynum, T.S. (2006) Fear of crime and criminal victimization: Gender-based contrasts. Journal of Criminal Justice 34(3): 285–301. Sloan, J.J. and Fisher, B.S. (2011) The Dark Side of the Ivory Tower: Campus Crime as a Social Problem. Cambridge, UK: Cambridge University Press. Spitzberg, B.H. and Hoobler, G. (2002) Cyberstalking and the technologies of interpersonal terrorism. New Media & Society 4(1): 71–92. U.S. Department of Education, National Center for Education Statistics. (2012) Digest of Education Statistics, 2011. Washington, DC: (NCES 2012-001), Chapter 3. Wilcox, P., Skubak Tillyer, M. and Fisher, B.S. (2009) Gendered opportunity?: School-based adolescent victimization. Journal of Research in Crime and Delinquency 46(2): 245–269.

330

© 2013 Macmillan Publishers Ltd. 0955-1662

Security Journal

Vol. 26, 4, 315–330