Risky Social Media Behaviors and the Potential for Victimization: A Descriptive Look at College Students Victimized by Someone Met Online

Abstract

This work uses original survey research to explore the nature of risky behaviors and exposure to potential offenders using social media as a means to initiate personal relationships. During fall 2016, instructors at a large university were recruited to conduct paper and pencil classroom surveys, resulting in a sample of 1310 undergraduate students. Their results first suggest that nearly all sampled students use social media and that half of college students use social media to meet new people. Alarmingly, ∼1 in 10 students disclosed personal information about themselves within 24 h of making initial contact with someone online. Furthermore, 1 in 20 young people chose to meet in person within 24 h of first contact. Among the survey respondents over 10% reported being a victim of stalking, cyberstalking, or sexual victimization resulting from a perpetrator they initially met on social media. The authors discuss these findings through a lifestyle-routine activity theory lens and explore the implications for understanding victimization risk.

Introduction

Over the last decade, and since the introduction of Facebook in 2005, social networking sites have rapidly grown in both quantity and ease of use, attracting a growing number of users. As of 2016, almost 90% of young adults, age 18–29, had an active Facebook account, with over half checking their accounts multiple times a day (Greenwood et al. 2016). Social networking and media applications have quickly become a mechanism for young people to maintain old relationships, pursue new ones, or privately become immersed in the lives of other active users. The move to increased online interconnectivity has arguably fostered an environment where individuals can engage in harmful or antisocial behaviors, such as online harassment, or using Facebook to engage in “excessive viewing of a particular user's profile and pictures, as well as repeatedly messaging or posting comments,” referred to in pop culture as “Facebook stalking” (Techopedia 2017). Arguably, the advent of social media has had the unintended consequence of creating new pathways to victimization and opportunities for perpetration. This work will explore the possible correlations between social media use and certain types of victimization, with a goal of providing future researchers focal areas to test a causal relationship.

Newer, mobile friendly, and less complex social media applications, such as Twitter and Instagram, enable perpetration and facilitate risky behavior by users. Very little personal information is required to create an account, thus nearly anonymous users may “follow” whomever they choose without approval. For example, Instagram is a location-based photo sharing application, which does not require mutual consent between users. Interestingly, in posting photos to one's account, users may share their physical location with all of their followers, many of whom may be unsolicited strangers. Thus, the user is unintentionally increasing their exposure to potential stalking offenders while not being able to control access to who sees their information. While there are likely many pathways to stalking victimization, cyberstalking, and sexual assault, this uncontrolled exposure in a social media setting requires empirical examination.

Recently, this topic was the subject of a motion picture entitled Ingrid Goes West (2017), where a young adult woman, Ingrid, becomes obsessed with the Instagram account of a lifestyle blogger. Ingrid eventually becomes so consumed that she travels across the country, tracking down her idol using only Instagram location posts, in an attempt to orchestrate a series of “casual” meetings. While the example may seem outlandish, it is becoming increasingly possible for inadvertent online behaviors to increase potential risk for victimization, both online and off.

Online interpersonal victimization has been labeled “quick and easy,” possibly due to the multiple options for contact through different applications (Greenwood et al. 2016; Melander 2000, p. 266). Among individuals 18–24, the majority have reported some form of online harassment, with women more commonly receiving sexually-based comments or becoming victim to cyberstalking behaviors (Duggan et al. 2015). Cyberstalking, often predicated by online harassment, includes threatening, unwanted, and persistent behaviors directed at a single individual (Cavezza and McEwan 2014; Reyns et al. 2012).

A recent German study of social networking platforms reported that almost half of users experienced online harassment, resulting in approximately half of these cases escalating to cyberstalking, with females being significantly more likely to be victims (Dreßing et al. 2014). In many instances, cyberstalking behaviors occur concurrently with traditional offline stalking behaviors (Dreßing et al. 2014; Nobles et al. 2014).

A Malaysian study of adolescents revealed that being approached through social networking applications for an in-person meeting was a relatively common experience, with 5.5% experiencing sexual victimization as a result of the meeting (Marrett and Choo 2018). Other studies have explored risk and exposure to victimization formed through online dating applications. Most notably, the United Kingdom conducted a nationwide study of sexual assault facilitated by online dating platforms where victims were overwhelming young adult females (ages 20–24). The study found a six-fold increase, from 33 offenses in 2009 to 184 in 2014 (National Crime Agency 2016). Unfortunately, research examining the connection between victimization and social media application use among the young adult population, an already high-risk victimization group with an increasing dependence on technology, is not readily available using a sample from the United States (Reyns et al. 2011).

Theoretical framework

In this study the authors examine the relationship between potentially risky social media behaviors and victimization risk through the lens of lifestyle-routine activity theory (L-RAT) (Cohen and Felson 1979; Hindelang et al. 1978). L-RAT suggests that individual risk of victimization varies by individual lifestyle choices that alter exposure to potential offenders, as well as determine attractiveness to a potential predator. For example, a college-age individual is likely to spend more of their time out socializing. These individuals are thus exposing themselves to more potential offenders.

Interestingly, and more to the point, the theory has been linked to online behavior as well. Rapid increase in online activity creates new pathways to victimization, as alluded to by the original theorists, “Many technological advancements designed for legitimate purposes…may enable offenders to carry out their own work more effectively” (Cohen and Felson 1979, p. 591). The key components of L-RAT have been linked to increased risk of cybervictimization through more time spent online (Bossler et al. 2012; Brady et al. 2016; Marcum et al. 2010; Ngo and Paternoster 2011; Pratt et al. 2010; Reisig et al. 2009; Reyns et al. 2011) and more personal information displayed on social media (Reyns et al. 2011), among others, with research in this area most commonly examining the relationship between online risky behaviors and online victimization.

The current study explores potentially risky social media behaviors through an application of L-RAT concepts such as exposure to potential offenders, target attractiveness, and lack of guardianship. The authors pay particular attention to the process of moving relationships from online to offline and, specifically, the mechanisms used by young adults to initiate offline meetings through online social media platforms. Furthermore, gender differences are explored based on theoretical propositions that males are more likely to engage in risky behaviors (Hindelang et al. 1978), yet females are more likely to experience interpersonal victimization (Belknap 2015). As contemporary youths and young adults become increasingly reliant and comfortable with technology-facilitated relationships, it is imperative to develop a clear picture of the potential risks.

Materials and Methods

Data were collected through paper and pencil surveys distributed to a random sample of undergraduate classrooms on a university with a total enrollment near 20,000. Classrooms were selected using a systematic random sample of the course offerings during the selected sampling period. A list of all fall 2016 course offerings was compiled through the university registrar, resulting in a total of 1750 classes. After calculating an average number of students per course, the ideal sample size for these data,¹ and expected instructor response rate, 150 classes were selected for solicitation emails by selecting every 18th class in the list after choosing a random starting point. The authors asked individual instructors for permission to distribute surveys and were ultimately given access to 47 different classrooms (31.3% of the classrooms solicited) across 28 university departments. The initial sample resulted in surveys collected from 1445 consenting students. Less than 5% of students declined to participate in the survey. Survey materials were approved by the university institutional review board before sampling. Based on this study's exclusive focus on younger and “traditional” college students, all cases were removed where the respondent was over the age of 30. Surveys were also removed if the respondent did not respond to over 75% of the questionnaire, resulting in a sample of 1310 students.

Focused creation of a subsample consisting of self-identified social media users resulted in an effective sample of 1252 college students (95.57% of the original sample). A comparison of social media users and nonusers can be seen in Table 1. The authors define social media users to be individuals that have reported ever using social media. While recording an individual as a user if they had ever used social media may seem to cast a wide net, the authors found that 96.8% of those broadly defined as users had reported using social media at least once in the past 30 days. Social media is exemplified in this study in four different forms, Facebook, Instagram, Snapchat, and Twitter. A “user” only needs to be an account holder for one of the four to be considered a social media user. Four domains of social media may seem limited, but 95.6% of the initial sample indicated use of at least one of these. Reflectively, the authors feel that these four domains are highly inclusive and represent “mainstream” social media. Basic demographic measures collected from respondents included their age, race, sex, and sexual orientation and can be found in Table 1.

Table 1.

Descriptive Characteristics of Respondents (N = 1310)

	Social media users (N = 1252)		Nonusers (N = 58)
	Mean	SD	Mean	SD	T
Age (17–30)	20.79	2.30	20.83	2.56	0.108
	n	%	n	%	χ²
Sex
Male	487	38.90	20	34.48	0.456
Female	765	61.10	38	65.52
Race
White	634	50.64	25	43.10	9.644^*
African American	248	19.81	21	36.21
Hispanic	303	24.20	9	15.52
Other	67	5.35	3	5.17
Sexual orientation
Heterosexual	1133	90.50	52	89.66	0.34
Sexual minority	119	9.50	5	8.62	0.34

p < 0.05.

Victimization variables

The focal point for this analysis is victimization as a direct result of a face-to-face encounter with an individual met initially through social media. The authors argue here that this encounter may represent a specific exposure to risk of victimization that warrants further exploration. The specific measures of victimization assessed in the sample include stalking, cyberstalking, and sexual assault. Stalking is measured as two or more positive responses to a seven-item series of questions about stalking victimization, including following/spying, “waiting outside or inside places,” and “leaving unwanted items, presents, or flowers,” among others. This series is notable as it is the indicator of stalking victimization used by the National Crime Victimization Survey's (NCVS) Supplemental Victimization Survey. To be classified as a stalking victim, respondents had to report experiencing two or more stalking behaviors as there was not a measure of “repeated” behaviors in the survey instrument. These items are commonly used in the current stalking victimization literature [see Fox et al. (2016); Nobles et al. (2014)] and provide a widely accepted basis for estimating stalking victimization in the sample.

Cyberstalking is measured using responses to a multi-item scale designed to address cyberstalking victimization [see Reyns et al. (2012)]. This scale consists of a number of dichotomous items where each of the items is designed with a specific unwanted behavior and a repeated process in mind, such as unwanted sexual advances or threats of violence, among others. For cyberstalking victimization, a respondent only need to indicate that they have repeatedly experienced one of the unwanted behaviors listed.

Finally, the authors asked respondents to complete the Sexual Experiences Survey (Koss et al. 1987). This survey consists of 10 dichotomous items that capture a range of nonconsensual behaviors and experiences, ranging from sexual coercion to rape, that have been demonstrated to be a valid estimate of sexual assault. The authors chose to use this version of the Sexual Experience Survey to address sexual assault in their sample both for its brevity and acceptance as a reliable instrument. If a respondent reported a “yes” for any of the 10 items, they were considered to be a victim of sexual assault.

Social media

Social media user behavior measures include frequency of use, disclosure of personal information, and importantly a number of details regarding any meetings that occurred in person. In part, due to a lack of existing validated scales and the exploratory nature of this research, all social media disclosure and behavior questions were developed by the researchers with specific measurement objectives. To capture the frequency of social media use the authors asked respondents a Likert type item where respondents rated how frequently they use social media. Response categories included less than once a week, once a week, once every few days, once a day, and multiple times a day. With respect to the specific risky behaviors of disclosure of personal information and exposure to potential offenders, the authors operationalized disclosure as providing a new contact with a personal phone number, a physical address, and/or exposure as the agreement to meet in person. To allow for those individuals who have either disclosed information or met in person, the authors asked respondents to provide information on what they consider “typical” of their experiences and behaviors. Respondents were asked to report on “typical” behaviors to get a general understanding of their social media behaviors and to avoid adding excess length to the survey. This strategy was deemed appropriate for the exploratory nature of this study and allowed for an exhaustive list of options for each item. These “typical” actions were first measured by whether or not an individual had ever agreed to any of the potentially risky behaviors and second based on the length of time they “usually” waited before sharing this information or agreeing to an in-person meeting.

Finally, the authors further address the nature of risk by ascertaining where and when individuals chose to meet. Options included a number of public places and private residences. Respondents were asked to report their “typical” meeting behaviors to accommodate those individuals who have met more than one person face-to-face after making an initial contact through social media. If respondents chose to meet in person, guardianship was measured by asking respondents if they took any protective measures before the meeting. These included physical protection, such as carrying a gun or pepper spray, making sure they had a fully charged cell phone, alerting friends or family of their plans, and/or posting their plans on social media. Respondents could select multiple options, resulting in an additive measure of protective behaviors ranging from 0 to 4.

Results

General application use

As stated above, of the 1310 initial survey respondents, the vast majority (n = 1252: 95.93%) reported a history of social media use, with 92.52% of users (n = 1212) active in the last month. The information presented in this section will focus on the subsample of 1252 respondents who reported any history of social media use. Overall, the preferred social media platform of the subsample was Facebook (95.93%), followed by Snapchat (88.98%), Instagram (73.40%), and Twitter (67.57%). Application use by sex is presented in Table 2. There were no sex differences for Facebook or Instagram use; however, female respondents were more likely to use Snapchat (χ² = 9.27, p < 0.01) and Twitter (χ² = 6.46, p < 0.01). The average user was active on three applications in the past month, with 37.25% (n = 488) of respondents reporting being active on all four applications within the last month.

Table 2.

Social Media Application Use by Sex

	Full user sample (N = 1252)		Males only (n = 487)		Females only (n = 765)		Comparison by sex
	n	%	n	%	n	%	χ²
Application use ever
Facebook	1201	95.93	465	95.48	736	96.21	0.01
Instagram	919	73.40	349	71.66	571	74.64	0.91
Snapchat	1114	88.98	412	84.60	702	91.76	9.27^**
Twitter	846	67.57	306	62.83	540	70.59	6.46^**
Frequency of checking account
<Once a week	62	4.95	31	6.37	31	4.05	8.63
Once a week	25	2.00	12	2.46	13	1.70
Once every few days	73	5.82	32	6.57	41	5.36
Once a day	198	15.81	83	17.04	115	15.03
Multiple times a day	924	73.80	333	68.38	591	77.25

p < 0.01.

As seen in Table 2, it was most common for respondents to check their accounts multiple times per day, with 73.80% (n = 924) of individuals engaging in this activity and female respondents checking their accounts more often overall (χ² = 8.63, p < 0.01). This finding provides support for a growing fixation with social media and technology-based communications.

Disclosure of personal information

As suggested by Reyns and colleagues (2011), disclosure of personal information through social media is positively correlated with risk of cybervictimization. For this study, disclosure was measured as whether or not a user shared their phone number or physical address with an online contact, how soon this occurred, and whether an in-person meeting took place. Results for the sample of social media users are presented in Table 3.

Table 3.

Time to Disclosure of Personal Information (N = 1252)

	Full user sample (N = 1252)		Males only (n = 487)		Females only (n = 765)		Comparison by sex
	N	%	n	%	n	%	χ²
Phone no. given	449	35.86	197	40.45	252	32.94	37.62^***
<1 h	43	9.58	26	5.34	17	2.22
1–12 h	76	16.93	44	9.03	32	4.18
12–24 h	51	11.36	29	5.95	22	2.88
1–2 days	72	16.04	38	7.80	34	4.44
Within 1 week	109	24.28	37	7.60	72	9.41
>1 Week	98	21.83	23	4.72	75	9.80
Phone no. NOT given	803	64.14	290	59.55	513	67.06
Address given	49	3.91	32	6.57	17	2.22	11.51^**
<1 h	4	8.16	4	0.82	0	0.00
1–12 h	4	8.16	4	0.82	0	0.00
12–24 h	2	4.08	2	0.41	0	0.00
1–2 days	7	14.29	4	0.82	3	0.39
Within 1 week	11	22.45	9	1.85	2	0.26
>1 Week	21	42.86	9	1.85	12	1.57
Address NOT given	1203	96.09	455	93.43	748	97.78
Meeting occurred	585	46.73	256	52.57	329	43.01
<1 h	16	2.74	12	2.46	4	0.52
1–12 h	22	3.76	10	2.05	12	1.57
12–24 h	28	4.79	19	3.90	9	1.18
1–2 days	53	9.06	40	8.21	13	1.70	101.63^***
Within 1 week	100	17.09	64	13.14	36	4.71
>1 Week	234	40.00	75	15.40	159	20.78
Meeting did NOT occur	667	53.27	231	47.43	436	56.99

Percentages for length of time are based on the number of respondents who reported engaging in this activity and not the full sample of users. Of the 585 respondents reporting that they had met someone from social media, 132 (22.56%) did not indicate the length of time from exchange of the first message to when the meeting occurred.

p < 0.01; ^***p < 0.001.

Generally, an in-person meeting facilitated through social media was more common than a user sharing their phone number through the medium (n = 585, 46.73% vs. n = 449, 35.86%, respectively). The least common risky behavior was a user sharing their physical address with an online-only contact (n = 49, 3.91%). Over half of male users (n = 256, 52.57%) reported meeting a social media contact in person, compared to 43% of female users (n = 329; χ² = 16.55, p < 0.01). Male users also reported sharing their phone number at a higher rate than female users (41.07% vs. 33.33%, respectively; χ² = 12.75, p < 0.01), as well as their physical address (6.37% vs. 2.22%, respectively; χ² = 15.15, p < 0.01). Furthermore, males reported meeting faster than female users (χ² = 101.63, p < 0.01), with the average meeting for males occurring within 1–2 days to a week of the first message exchanged, compared to within or more than 1 week for female users.

Meeting behaviors

Male users reported meeting a greater number of their online-only contacts in a face-to-face setting with almost 17% of male users (n = 84) reportedly meeting over six social media contacts in person, compared to 8% (n = 58) of female users (χ² = 40.44, p < 0.01). These differences by sex are displayed in Table 4.

Table 4.

Social Media Meeting Behaviors by Sex

	Full user sample (N = 1252)		Males only (n = 487)		Females only (n = 765)		Comparison by sex
	N	%	n	%	n	%	χ²
No. of people met through social media
0	484	38.66	153	31.42	331	43.27	40.44^**
1 to 2	281	22.44	98	20.12	183	23.92
3 to 5	162	12.91	74	15.20	88	11.50
6 or more	142	11.58	84	17.25	58	7.58

Respondents who have met someone from a social media application
	Full sample		Males only		Females only
	(n = 585)		(n = 256)		(n = 329)
Typical meeting time
Before 6:00 PM	233	39.83	99	38.67	134	40.73	0.25
6:00–9:00 PM	217	37.09	109	41.80	108	32.84	5.87^*
After 9:00 PM	86	14.70	49	19.14	37	11.25	7.16^**
Typical meeting place
Public with alcohol	180	30.77	94	36.72	86	26.14	7.56^**
Public without alcohol	170	29.06	63	24.61	107	32.52	4.37^*
Private residence	194	33.16	106	41.41	88	26.75
Respondent's home	48	8.21	38	14.84	10	3.04	26.64^**
Other user's home	64	10.94	44	17.19	20	6.08	18.23^**
Third party home	133	22.74	65	25.39	68	20.67	1.83
Precautions taken
Fully charged phone	408	69.74	179	69.92	229	69.60	0.07
Physical protection	180	30.77	47	18.36	133	40.43	32.9^**
Tell family/friends	385	65.81	151	58.98	234	71.12	9.43^**
Post on social media	61	10.43	18	7.03	43	13.07	5.62^*

Percentages may equal more than 100% due to respondents selecting multiple answers.

p < 0.05; ^**p < 0.01.

Most commonly, meetings occurred before 6:00 PM (n = 233, 39.83%), followed by those between 6:00 and 9:00 PM (n = 217, 37.09%). Meetings past 9:00 PM were least common in this sample (n = 86, 14.70%) with males more likely to engage in these late-night meetings compared to females (χ² = 7.16, p < 0.01).

Meeting location is also presented in Table 4. Interestingly, the most common place to meet an online contact was at a private residence (n = 194, 33.16%), whether that be the respondent's residence, the other user, or a third party. Meeting at a public place with alcohol was the second most common (n = 180, 30.77%) followed by a public place without alcohol (n = 170, 29.06%). Arguably, it is the riskier locations that were more frequently chosen for meeting locations among respondents in this sample. When examining meeting place by sex, males were more likely to meet new contacts in public places with alcohol (χ² = 7.56, p < 0.01), at their own residence (χ² = 26.64, p < 0.01), or at the residence of the other user (χ² = 18.23, p < 0.01). Female users were more likely to meet in a public place without alcohol (χ² = 4.37, p < 0.01).

To measure guardianship during meetings, respondents were asked to report on the typical protections they took, including having a fully charged cell phone, carrying some form of physical protection (i.e., pepper spray), informing family or friends about their plans, or posting their plans on social media. These results, separated by sex, are displayed in Table 4. On average, female respondents took more precautions than male respondents (M = 2.55 vs. M = 1.73, respectively), with females most commonly sharing their plans with family/friends (n = 234, 71.17%) and males most commonly carrying a fully charged cell phone (n = 179, 69.92%).

Victimization

Victimization in this study was measured through responses to a series of items that were operationalized as stalking, cyberstalking, or sexual assault. Respondents may not have reported the behavior to law enforcement or even considered the incident to be a form of victimization. In this study, ∼30% (n = 375) of respondents reported some history of sexual assault. Approximately one-third (n = 410) of social media users reported experiencing stalking victimization. Finally, 43% (n = 545) of respondents reported some form of cyberstalking victimization. Reported victimization may be found in Table 5.

Table 5.

Victimization by Sex

	Full user sample (N = 1252)		Males only (n = 487)		Females only (n = 765)		Comparison by sex
	n	%	n	%	n	%	χ²
Any victimization	711	56.79	220	45.17	491	64.18	32.96^**
Stalking	410	32.75	128	26.28	282	36.86	15.12^**
Cyberstalking	545	43.53	160	32.85	385	50.33	30.97^**
Sexual Assault	375	29.95	83	17.04	292	38.17	53.85^**

Respondents who have met someone from a social media application
	(n = 585)		(n = 256)		(n = 329)
From a social media contact exclusively	74	12.65	23	8.98	51	15.50
Stalking	42	7.18	16	6.25	26	7.90
Cyberstalking	64	10.94	20	7.81	44	13.37
Sexual assault	12	2.05	3	1.17	9	2.74

p < 0.01.

Overall, of the 1252 social media users in their sample, 711 (56.79%) reported experiencing some form of victimization, whether sexual assault, traditional stalking, or cyberstalking. Of those individuals reporting victimization, ∼1 in 10 (n = 74, 12.65%) noted an offender that the victim initially met through a social media application. The most common form of victimization from a social media offender was cyberstalking, with 64 respondents experiencing some form of these behaviors, the vast majority of victims being female (n = 44, 68.75% of cyberstalking victims with a social media offender). Traditional in-person stalking was the second most common form of victimization among social media users, with 42 users reporting experiencing these behaviors from a social media offender. Stalking victimization was also more common among female respondents (n = 26, 59.09% of stalking victims with a social media offender). Forty respondents reported experiencing both cyberstalking and stalking behaviors perpetrated by a social media acquaintance.

Finally, 12 respondents reported experiencing sexual victimization facilitated through social media, with nine female and three male victims. More simply put, of the users who participated in face-to-face meetings with social media contacts, ∼1 in 50 respondents reported some form of sexual assault victimization from an offender first met through social media.

Discussion and Conclusion

This work has been exploratory and descriptive in nature with the intent to provide some basic information on the relationship between risk and the online to face-to-face transitional behavior of college students and victimization of college students. Overall, approximately one in three respondents reported a history of sexual assault. Past research using the Sexual Experiences Scale has found similarly high levels of victimization (Humphrey and White 2000), with reports as high as 50% (Koss et al. 1987). Due to the measurement of the stalking variable, as two or more behaviors, it can be suggested that stalking victimization is underestimated in this study as it does not include individuals who have experienced the same behavior multiple times (i.e., a victim who has been repeatedly harassed using telephone calls or text messages only). Regardless, one in three respondents reported experiencing some form of stalking victimization. Estimates of traditional stalking vary widely, from 2% in a national sample (Catalano 2012) to 26% in a sample of college students (Fox et al. 2016), with higher percentages reported for female respondents. Finally, cyberstalking victimization rates reported, 43%, were similar to the original study from which cyberstalking measures were derived (Reyns et al. 2012).

The results of this exploration suggest that social media can act as a catalyst for young people looking to meet and make new acquaintances. In contrast, 74 individuals or 5.9% (12.65% of those who met in person) of the sample of social media users reported some form of victimization perpetrated by an offender they met through social media. The majority of these victimizations were either stalking or cyberstalking, but there were also 12 reported sexual assaults. This sample is representative of the full university population it was taken from and as such the authors can extrapolate these findings to that larger group. For example, 46.7% of the sample respondents met someone in person that they had initially made contact with through a social media application. Approximately 13% of those reported some form of stalking, cyberstalking, or sexual assault. On a campus of roughly 20,000 students this is over 2500 victims of stalking, cyberstalking, or sexual assault facilitated through social media applications.

This work highlights a number of significant correlations between victimization and certain types of online behaviors, and while these cannot demonstrate causation, the authors have taken an important step by demonstrating that an empirical relationship does exist. Theoretically, the authors believe that the process of disclosing personal information and meeting face-to-face, under certain circumstances, represents a pathway to victimization [similar to those discussed in Turanovic et al. (2015)]. In addition, it would seem that social media is an entry point to that pathway. Social media alone does not present a risk to the user, insofar as the risk is a product of the way the user utilizes social media. It has been suggested that including personal information in online spaces is an inherently risky behavior [e.g., Brady et al. (2016)]. Likewise, while disclosing information on social media is a risk, it is not determinable here if that risk is any greater than disclosing personal information in any other setting. Rather, it may be a rush to meet and mismatched expectations that are the source of victimization. Out of 585 respondents that chose to meet in person, roughly 11% (n = 66) did so within 24 h, nearly 40% of those who disclosed a phone number did so in 24 h or less, and roughly 20% (n = 10) of those who disclosed a physical address did so in 24 h or less. The pursuit of immediate gratification and its resemblance to low self-control may account for some of the victimization the authors see in this study. Research in this area consistently finds links between measured low self-control and victimization, and it is suggested that lower self-control is directly related to risk management [e.g., Schreck (1999)]. Thus, it may be that these data are merely highlighting the connection between self-control and victimization risk.

Male and female respondents present an interesting difference in this sample that may constitute a microcosm of gendered behavior and how closely related exposure and predations are. Males disclose more information more frequently, are more likely to meet in person, and meet more than one or two persons. Males are more likely to meet in places which serve alcohol or at a private residence and take far fewer precautions (Table 4). Males are also considerably less likely to be a victim of stalking, cyberstalking, or sexual assault. Rather, males are far more likely to be a perpetrator of these crimes, and thus, their “risky” behavior may be less about exposure and vulnerability and more about exhibiting low self-control and creating opportunity. Unfortunately, perpetration was not measured in this study, but would be a potential area for future research.

The “Lifestyle/Routine Activity” perspective suggests that victimization is ultimately a product of exposure to risk (Cohen et al. 1981; Hindelang et al. 1978). Prior research has supported this notion with correlations between risk and victimization for female respondents. Risky behaviors are often long term and persistent, leading to a higher chance of repeat victimization among these individuals (Averdijk 2011; Fisher et al. 2010). Even after an individual has experienced a victimization event, they are not likely to make broad behavior changes (Averdijk 2011). Therefore, even though female respondents in this sample engaged in less risky behaviors than males, it can be suggested that the victims in this sample will have a higher risk of later victimization compared to their nonvictimized counterparts or may have already experienced multiple victimization events.

What remains uncertain in this descriptive work is to what degree social media itself represents exposure or risk. However, what is clear is that a significant proportion of college students are using social media as a mechanism for meeting new people and forming offline relationships. It is also possible that social media is being used as a conduit for forming and maintaining interpersonal relationships. It would also seem here that when social media is being used to initiate and pursue sexual relationships, and when individuals are quick to disclose personal information or meet, the process rather than the tool represents an inherently risky behavior.

Limitations

In any study of this nature there are limitations that should be weighed when attempting to draw conclusions from the work. In the work presented in this study the authors must acknowledge certain key limitations that are the result of the method and compromises made to ensure that the data would be collected. First, these data are cross-sectional, which presents a number of inherent limitations even to descriptive work. Thus, to use the term “pathway” may be presumptive as there are elements to the correlation between risk, disclosure, exposure, and social media use that cannot be disentangled here. To accommodate some of these limitations, principally causal ordering, the authors did ask respondents to refer to victimizations where they knew the perpetrator and that individual was someone that they had met through social media. Even so, the authors cannot be certain that the specific event is a result of their described risky behaviors, as those behaviors are aggregated over time. However, the existence of the reported correlations is suggestive and warrants further research. To demonstrate causality, future research should direct specific survey items to address victimization events that occurred as a result of social media interactions and ask respondents to describe the behaviors and activities leading up to the incidents. Longitudinal and panel research could also be used to examine risky social media behaviors and victimization. Finally, the authors did not assess victimization among nonsocial media users due to the small size of this subsample. The authors would need a much higher sample of nonusers, which may not be realistic given the current state of technology and reliance on social media accounts. The reality is more likely that social media itself is relatively harmless, but certain types of behavior on these platforms constitute considerable risk.

Second, the authors elected to use the 1987 version of the Sexual Experiences Survey, rather than an updated version [see Koss et al. (2007)]. The authors did so to limit the total time that it would take to complete the larger survey instrument. While adding the newer version of the SES would have only added 25 more items, the total instrument was already 19 pages and required many in the pilot groups over 30 min to complete. The authors had concerns that the additional questions would result in large amounts of missing data.

Finally, in electing to use the National Crime Victimization Survey questionnaire on stalking the authors were using an imperfect measure of stalking. The nature of stalking is that the behavior is repeated over time, and the instrument does not accommodate this dimension well.

Implications and conclusion

Interestingly, the inclusion of social media elements of this study is the result of multiple rounds of piloting the instrument. In part, the instrument was meant to capture the role of mobile dating applications as an exposure to risk of victimization. The focus was on mobile applications specifically designed to bring people together from online to face-to-face (e.g., Tinder, Bumble, and so on). The question and answer sessions during piloting brought to their attention a need for addressing contact initiated through a social media platform. If not for these suggestions, the authors likely would not have thought of social media as a facilitator of new relationships.

Future research needs to address a number of interesting corollaries found in this study. Principally, future work needs to explore whether or not this pathway is indeed riskier than “typical” college student behavior. An important aspect of future research includes a survey revision and collection of more data. Due to the exploratory nature of this study, many of the items were designed to cover all possible outcomes and possibilities. Based on the initial data, subsequent surveys should collect more data on individuals that would appear most “at-risk” for victimization, such as those individuals who typically use online applications to pursue quick late-night meetings at private residences. Furthermore, these behaviors should be explored in relation to sexual orientation or certain attitude and behavior variables, such as self-control and self-esteem.

Continued research in this area is needed to inform criminal justice agencies about the new possibilities for victimization risk and suggestions for response. Currently, agencies struggle with who is responsible for pursuing and punishing offenders through online applications, most often with regards to online harassment. It often becomes a back-and-forth between local law enforcement or the application corporation, with little being accomplished for the victim (Hess 2014). Furthermore, individuals victimized through these means may simply be told to “log out” or “deactivate” their applications, which places responsibility in the hands of the victim, a response that may be unique to this form of crime (Powell and Henry 2016).

Furthermore, stalking and/or cyberstalking may be occurring without the victim's awareness. As in the case of Ingrid, the character engaged in unhealthy behaviors that began online and manifested into an offline context of stalking, with the victim being generally unaware of what was transpiring. Popular social media applications are becoming places where it is not unusual for interactions between strangers to begin online and move to an offline meeting, which is in stark contrast to the traditional formation of platonic or romantic relationships. While it may seem that individuals who desire to seek relationships through online means have many options in the world of online dating applications, it can be suggested that social media provides an increased feeling of security and decreased feeling of sexual expectations for one or both parties, especially compared to “hook-up” apps, such as Tinder.

Footnotes

Author Disclosure Statement

No external funding or conflicts of interest existed for this project.

References

Averdijk

. (2011). Reciprocal effects of victimization and routine activities. J Quant Criminol. 27, 125–149.

Belknap

. (2015). The Invisible Woman: Gender, Crime, and Justice (4th edition). (Cengage Learning). Stamford, CT.

Bossler

, Holt

, May

. (2012). Predicting online harassment victimization among a juvenile population. Youth Soc. 44, 500–523.

Brady

, Randa

, Reyns

. (2016). From WWII to the World Wide Web: A research note on social changes, online “places,”. and a new online activity ratio for routine activity theory. J Contemp Crim Justice. 32, 129–147.

Catalano

. (2102). Stalking victims in the United States–Revisited. (Publication No. NCJ 224527). Washington, DC: Bureau of Justice Statistics.

Cavezza

, McEwan

. (2014). Cyberstalking versus off-line stalking in a forensic sample. Psychol Crime Law. 20, 955–970.

Cohen

, Felson

. (1979). Social change and crime rate trends: A routine activity approach. Am Sociol Rev. 44, 588–608.

Cohen

, Kluegel

, Land

. (1981). Social inequality and predatory criminal victimization: An exposition and test of a formal theory. Am Sociol Rev. 46, 505.

Dreßing

, Bailer

, Anders

, et al. (2014). Cyberstalking in a large sample of social network users: Prevalence, characteristics, and impact upon victims. Cyberpsychol Behav Soc Netw. 17, 61–67.

10.

Duggan

, Ellison

, Lampe

, et al. (2015). Online Harassment. (Pew Research Center, Washington, DC) pp. 1–65. Retrieved from http://www.centralvalleybusinesstimes.com/links/PI_OnlineHarassment_102214-1.pdf (accessed May, 2018).

11.

Fisher

, Daigle

, Cullen

. (2010). What distinguishes single from recurrent sexual victims?. The role of Lifestyle‐Routine Activities and first‐ incident characteristics. Justice Q. 27, 102–129.

12.

Fox

, Nobles

, Fisher

. (2016). A multi-theoretical framework to assess gendered stalking victimization: The utility of self-control, social learning, and control balance theories. Justice Q. 33, 319–347.

13.

Greenwood

, Perrin

, Duggan

. (2016). Social Media Update 2016: Numbers, Facts and Trends Shaping the World. (Pew Research Center, Washington, DC) (pp. 1–19).

14.

Hess

. (2014). Why Women Aren't Welcome on the Internet. (Pacific Standard) p. 6. Retrieved from. https://psmag.com/why-women-aren-t-welcome-on-the-internetaa21fdbc8d6#.4am3xg29s (accessed May, 2018).

15.

Hindelang

, Gottfredson

, Garofalo

. (1978). Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. (Ballinger, Cambridge, MA).

16.

Humphrey

, White

, (2000). Women's vulnerability to sexual assault from adolescence to young adulthood. J Adolesc Health, 27, 419–424.

17.

Koss

, Gidycz

, Wisniewski

. (1987). The scope of rape: Incidence and prevalence of sexual aggression and victimization in a national sample of higher education students. J Consult Clin Psychol. 55, 162.

18.

Marcum

, Higgins

, Ricketts

. (2010). Potential factors of online victimization of youth: An examination of adolescent online behaviors utilizing routine activity theory. Deviant Behav. 31, 381–410.

19.

Marret

, Choo

W-Y

. (2018). Victimization after meeting with online acquaintances: A cross-sectional survey of adolescents in Malaysia. J Interpers Violence. 33, 2352–2378.

20.

Melander

. (2010). College students' perceptions of intimate partner cyber harassment. Cyberpsychol Behav Soc Netw. 13, 263–268.

21.

National Crime Agency. (2016). Emerging New Threat in Online Dating: Initial Trends in Internet Dating-Initated Serious Sexual Assaults. (National Crime Agency, United Kingdom).

22.

Ngo

, Paternoster

. (2011). Cybercrime victimization: An examination of individual and situational level factors. Int J Cyber Criminol. 5, 773.

23.

Nobles

, Reyns

, Fox

, Fisher

. (2014). Protection against pursuit: A conceptual and empirical comparison of cyberstalking and stalking victimization among a national sample. Justice Q. 31, 986–1014.

24.

Powell

, Henry

. (2016). Policing technology-facilitated sexual violence against adult victims: Police and service sector perspectives. Policing Soc. 28, 291–307.

25.

Pratt

, Holtfreter

, Reisig

. (2010). Routine online activity and internet fraud targeting: Extending the generality of routine activity theory. J Res Crime Delinq. 47, 267–296.

26.

Reisig

, Pratt

, Holtfreter

. (2009). Perceived risk of internet theft victimization: Examining the effects of social vulnerability and financial impulsivity. Crim Justice Behav. 36, 369–384.

27.

Reyns

, Henson

, Fisher

. (2011). Being pursued online: Applying cyberlifestyle–routine activities theory to cyberstalking victimization. Crim Justice Behav. 38, 1149–1169.

28.

Reyns

, Henson

, Fisher

. (2012). Stalking in the twilight zone: Extent ofcyberstalking victimization and offending among college students. Deviant Behav. 33, 1–25.

29.

Schreck

. (1999). Criminal victimization and low self-control: An extension and test of a general theory of crime. Justice Q. 16, 633–654.

30.

Spicer

. (Director). (2017). Ingrid goes west [Motion picture]. United States: Neon.

31.

Techopedia. (2017). Facebook stalking. Retrieved from https://www.techopedia.com/definition/27873/facebook-stalking

32.

Turanovic

, Reisig

, Pratt

. (2015). Risky lifestyles, low self-control, and violent victimization across gendered pathways to crime. J Quant Criminol. 31, 183–206.