Technology-Based Interpersonal Victimization

Abstract

The objective of this study was to identify factors that could predict youth’s future technology-based interpersonal victimization and the pattern of that future victimization over time. Data from Growing up With Media, a national, longitudinal, online study were analyzed. At baseline, participants (N = 1,018) were 10- to 15-year-old English speakers who had used the Internet at least once in the last 6 months. Twenty-nine percent reported repeat technology-based interpersonal victimization over a 2-year period (re-victimized group); 10% were victims during only Year 1 (desisted victimized group); and 17% reported victimization during only Year 2 (later victimized group). Of the individual risk factors examined, prior technology-based interpersonal victimization and current amount of Internet use had the strongest overall associations with pattern of technology-based interpersonal victimization over the subsequent 2-year period. There was substantial overlap among the individual risk factors. Thus, they could be thought of more simply in terms of four latent risk and three individual risk factors. On average, across these seven risk factors, repeat victims had the greatest average risk score (0.21) and the not victimized group had the lowest (−0.16). Repeat victims were more likely to be female and older and had more prior experience with problem behaviors, substance use, and negative parent–child relationships as compared with the other three groups. Being female, prior experience with problem behavior, prior substance use, and prior negative parent–child relationships were also associated with frequency of technology-based interpersonal victimization in the near (Year 1) and more distant (Year 2) future. Many of these risk factors related to technology-based victimization over time are malleable, suggesting opportunities for effective targeting of future prevention efforts.

Keywords

technology-based interpersonal victimization predicting factors

Costs associated with youth victimization are substantial (Finkelhor & Hashima, 2001). To reduce these costs, research that identifies factors that predict future victimization so that the potential trajectory can be affected before the victimization occurs is critical. As new technologies become increasingly influential in the lives of youth (Lenhart, Madden, Macgill, & Smith, 2007), it is crucial to broaden our discussions about youth victimization. We use the term technology to encompass experiences that occur via the Internet (i.e., online) and/or via text messaging. Conversely, non-technology-based, or traditional, victimization is a term used to reflect experiences that occur through other communication modes such as face-to-face. The current article investigates what prior factors are predictive of technology-based interpersonal victimization over time.

Interpersonal victimization can take multiple forms. Relational aggression is an intentional act that negatively affects an individual’s reputation or social status such as someone spreading rumors about another person. Physical aggression includes threats of physical harm. Youth also can be victimized by unwanted sexual solicitations, which involve unwanted requests to talk about sex, give personal sexual information, or do something sexual (Finkelhor, Mitchell, & Wolak, 2000). Each year, unwanted sexual solicitation affects 9% to 18% of youth via the Internet (Jones, Mitchell, & Finkelhor, 2012; Ybarra, Mitchell, & Korchmaros, 2011) and 10% of youth via text messaging (Ybarra et al., 2011). In addition, each year, between 11% and 39% of youth are victims of physical or relational aggression via the Internet (Jones et al., 2012; Ybarra et al., 2011) and 24% are such victims via text messaging (Ybarra et al., 2011).

Research has identified co-occurring factors that are associated with increased likelihood of technology-based sexual solicitation, relational aggression, and physical aggression. These factors are hypothesized to be associated with technology-based victimization because they reflect opportunities for victimization (see Routine Activities theory, Cohen & Felson, 1979), they are individual characteristics that somehow elicit aggression, or they reflect particulars of the interaction between the victim and the perpetrator that elicit victimization. Girls are more likely than boys to report unwanted sexual solicitations online (Mitchell, Finkelhor, & Wolak, 2007). Older youth are more likely than younger youth to report physical or relational victimization (Ybarra et al., 2011; Ybarra, Mitchell, Wolak, & Finkelhor, 2006) and online unwanted sexual solicitations (Mitchell et al., 2007). Such forms of interpersonal victimization are also related to how youth use the Internet (Ybarra et al., 2011). Online physical or relational victimization is more common among youth who perpetrate this aggression (Ybarra et al., 2006), as well as among youth who are victims of offline physical or relational aggression (Ybarra et al., 2006). Technology-based victimization has also been associated with poor caregiver–child relationships (e.g., Mitchell et al., 2007; Ybarra, Diener-West, & Leaf, 2007), substance use (Ybarra et al., 2007; Ybarra & Mitchell, 2004), poor school performance (Ybarra et al., 2007), offline aggressive behavior (Ybarra & Mitchell, 2007), and delinquency (Ybarra & Mitchell, 2004). Given the overlaps in characteristics, caregiver–child relationships, and behaviors of youth experiencing these three different types of technology-based interpersonal victimization, in combination with the health imperative to reduce prevalence rates of all types of victimization, we examine the prediction of these three victimization types together as a global interpersonal victimization experience.

Current understanding of predictive factors of technology-based victimization is limited by cross-sectional analyses that do not provide information about directionality (e.g., Jones et al., 2012; Williams & Guerra, 2007) nor speak directly to identifying risk for future victimization. Research into the risk factors associated with future technology-based interpersonal victimization is critical for increasing our capacity to identify youth who may be particularly at risk of these experiences. These factors might help to identify youth at risk for victimization in the near or the more distant future. For some youth, their risk might not actualize into victimization until the more distant future when a particular confluence of factors comes together. Furthermore, youth who are at risk for chronic victimization need special research attention because chronic victims are more likely to report psychological distress than their peers (e.g., Turner, Finkelhor, & Ormrod, 2006). There is a clear benefit to being able to interrupt this pattern of victimization (Spatz Widom, Czaja, & Dutton, 2008) by identifying malleable factors that portend risk not just for victimization but also for re-victimization.

Using longitudinal data from Growing up With Media (GuwM), we (a) examine the overall relationships between risk factors (i.e., as identified in cross-sectional research) at baseline and the pattern of future victimization experienced over the subsequent 2 years (i.e., not victimized during either year; victimized during both years; victimized in the first year only; and victimized in the second year only); and (b) examine the profiles of these four victimization groups across identified individual and latent risk factors. In addition, we examine the relationships between the identified risk factors and the frequency of technology-based interpersonal victimization (a) in the near future (experienced during the first subsequent year) and (b) in the more distant future (experienced during the second subsequent year).

Method

Participants

The final analytic sample included respondents who participated in all three survey waves: 1,018 male (50%) and female (50%) youth, ages 10 to 15 years old, who used the Internet at least once in the past 6 months. Additional sample characteristics are displayed in Table 1.

Table 1.

Respondent Descriptors at Year 0 of Unweighted and Weighted^a Analytic Sample^b of GuwM Respondents and of Weighted Excluded GuwM Respondents.^c

	Analytic Sample		Excluded Respondents
	n = 1,018	n = 1,020	n = 563
Respondent Descriptors	Unweighted %	Weighted %	Weighted %
Female	50	50	52
Hispanic ethnicity	11	16^d,e	22
Race
White	81	77	75
Black or African American	12	14	14
Mixed racial background	7	10	11
Used alcohol in past year	10	11^e	16
Used marijuana in past year	4	5	6
Currently do not have Internet household rules	5	6	5
Currently do not have blocking software on the computer	58	58	58
Honest when responding to the survey	97	97^e	93
Alone when responding to the survey	50	51	50
	Unweighted Mean	Weighted Mean	Weighted Mean
Age	12.46	12.55	12.54
Current academic problems	6.40	6.29^d	6.23
Current time spent on the Internet during typical day	3.41	3.41	3.38
Current lack of parental enforcement of household Internet rules	1.60	1.60	1.56
Current lack of general parental monitoring	2.84	2.85	2.93
Current lack of youth emotional bond with caregiver	−0.06	−0.00	0.00
Current coercive parental discipline	5.40	5.43	5.54
Delinquent behavior in past year	0.95	0.92^e	1.25
Technology-based perpetration of interpersonal aggression in past year	0.65	0.65^e	1.25
Technology-based interpersonal victimization in past year	1.43	1.48^e	2.19
Traditional perpetration of interpersonal aggression in past year	1.94	1.86^e	2.89
Traditional interpersonal victimization in past year	2.72	2.69^e	3.20

Note. GuwM = Growing up With Media; HPOL = Harris Poll Online.

Data were weighted statistically to reflect the population of adults with children ages 10 to 15 years old in the United States and to adjust for adult respondents’ self-selection into the Internet-using population and the HPOL.

Analytic sample included GuwM respondents who provided usable data for all 3 years (i.e., Year 0, Year 1, and Year 2).

Excluded GuwM respondents did not provide usable data for all 3 years.

Weighted analytic sample differed reliably from unweighted analytic sample.

Weighted analytic sample differed reliably from weighted sample of excluded GuwM respondents.

Measures

The GuwM methodology reports (GuwM, 2007, 2008, 2009) provide more detail about all the measures. Descriptive statistics for continuous variables are summarized in Table 2.

Table 2.

Item/Scale Descriptive Statistics for the Weighted^a Analytic Sample^b of GuwM Respondents.

Respondent Descriptors	M	SD	Range
Age	12.55	1.72	10-15
Current academic problems	6.29	1.37	1-8
Current time spent on the Internet during typical day	3.41	1.38	1-6
Current lack of parental enforcement of household Internet rules	1.60	0.96	1-4
Current lack of general parental monitoring	2.85	1.26	2-10
Current lack of youth emotional bond with caregiver	−0.00	2.32	−2.7-11.2
Current coercive parental discipline	5.43	1.42	2-10
Delinquent behavior in past year	0.92	1.18	0-9
Technology-based perpetration of interpersonal aggression in past year	0.65	2.08	0-45
Technology-based interpersonal victimization in past year	1.48	3.32	0-45
Traditional perpetration of interpersonal aggression in past year	1.86	2.89	0-55
Traditional interpersonal victimization in past year	2.69	2.96	0-25

Note. GuwM = Growing up With Media; HPOL = Harris Poll Online.

Analytic sample included GuwM respondents who provided useable data for all 3 years (i.e., Year 0, Year 1, and Year 2).

Predicted outcome: Pattern of technology-based interpersonal victimization over the subsequent 2-year period (Year 1-Year 2)

Technology-based interpersonal victimization during the last year was queried using nine items based on prior youth surveys (Centers for Disease Control and Prevention [CDC], 2005; Finkelhor et al., 2000; GuwM, 2008; Cronbach’s α: Y1 and Y2 = .89): (a) Someone sent a text message that said rude or mean things; (b) someone made a rude or mean comment to me online; (c) someone spread rumors about me online, whether they were true or not; (d) someone sent a text message that was sexual in any way when I did not want to receive it; (e) someone sent a picture text message that was sexual in any way when I did not want to receive it; (f) someone tried to get me to talk about sex online when I did not want to; (g) someone online asked me for sexual information about myself when I did not want to tell the person, for example, really personal questions, such as what my body looks like or sexual things I have done; (h) someone asked me to do something sexual when I was online that I did not want to do; and (i) someone made a threatening or aggressive comment to me online. Response options included never in the last 12 months, less than a few times a year, a few times a year, once or twice a month, once or twice a week, and everyday/almost every day.

For Year 1 and Year 2 separately, youth who reported being victimized in any of the queried technology-based ways in the past year were coded as victimized during that Year. We then classified those who were not victimized during either Year 1 or 2 as not victimized; those who were victimized at least once by at least one type of victimization during Years 1 and 2 as being re-victimized; those who were victimized during Year 1 but not Year 2 as desisted victimized; and those who were victimized during Year 2 but not Year 1 as later victimized.

Posited previous year (Year 0) predictive factors

Experience with problem behavior

Response options, unless otherwise specified, ranged from never (0) to everyday/almost every day (5). For all of these variables, summation variables were created with larger values reflecting more of the construct. Technology-based interpersonal victimization¹ during Year 0 was assessed with the 9 items described above (Cronbach’s α = .87; possible range = 0-45). Technology-based perpetration of interpersonal aggression¹ was assessed with nine items that mirrored the victimization items (CDC, 2005; Finkelhor et al., 2000; GuwM, 2008; Cronbach’s α = .90; possible range = 0-45): (a) Made rude or mean comments to anyone online; (b) spread rumors about someone online, whether they were true or not; (c) sent a text message that said rude or mean things; (d) made aggressive or threatening comments to anyone online; (e) tried to get someone to talk about sex online when they did not want to; (f) tried to get someone to do something sexual online when they did not want to; (g) asked someone online about really personal sexual information when they did not want to give it; (h) sent a text message that was sexual in any way when that person did not want to receive it; and (i) sent a picture or video text message that was sexual in any way when that person did not want to receive it. Traditional interpersonal victimization¹ was assessed with 5 items that queried past-year experiences of specific types of victimization (CDC, 2005; Hamby, Finkelhor, Turner, & Kracke, 2011; Cronbach’s α = .70; possible range = 0-25): (a) Someone my age did not let me in their group anymore because they were mad at me; (b) someone spread a rumor about me, whether it was true or not; (c) someone stole something from me—for example, a backpack or wallet; (d) another person or group attacked me; and (e) someone pulled a knife or gun on me. Traditional perpetration of interpersonal aggression¹ was assessed with 11 items that queried past-year behaviors (CDC, 2005; GuwM, 2007; Cronbach’s α = .84; possible range = 0-55): (a) Excluded someone from your group; (b) spread a rumor about someone; (c) shoved, or pushed, or hit or slapped another person your age; (d) threatened to hurt a teacher; (e) threatened someone with a weapon (gun, knife, club, etc.); (f) used a knife or gun or some other kind of weapon such as a bat to get something from someone else; (g) been in a fight in which someone including yourself was hit; (h) gotten into a fight where a group of your friends were against another group of people; (i) hurt someone badly enough that they needed to be treated by a doctor or nurse; (j) stabbed or shot someone; and (k) kissed, touched, or done anything sexual with another person when that person did not want you to. Delinquent behavior¹ was assessed with 9 items asking about specific delinquent behaviors enacted in the past year (CDC, 2005; Finkelhor et al., 2000; Cronbach’s α = .69): (a) Banged up or damaged something that did not belong to you; (b) started a fire on purpose, where you wanted something to get damaged or destroyed; (c) broken into someone else’s house, building, or car; (d) lied to someone to get something that you wanted, or to get someone to do you a favor, or to get out of doing something you didn’t want to do; (e) taken something that was valuable, such as shoplifting or using someone else’s credit card, when no one was looking; (f) hurt an animal on purpose, such as cutting off its tail, hitting or kicking it, or killing it for fun; (g) stayed out at night even though you knew your parents would not want you to; (h) run away from home and stayed away overnight; and (i) ditched/skipped school. Response options for 3 of the 9 items included at least once in the past 12 months and never. After recoding all items individually to reflect delinquent behavior in the past 12 months (1) or not (0), we created a summation variable (possible range = 0-9).

Substance use

Substance use was assessed with two items (CDC, 2006) that measured any use of alcohol or marijuana in the past year (yes [1] vs. no [0]).

Parent–child relationship characteristics

(Finkelhor et al., 2000) Youth’s current emotional bond with their caregiver was assessed with three items that asked about “the parent or guardian in the home who knows the most about [the youth]” (Cronbach’s α = .61): (a) How well would you say you and this person get along? (b) Do you feel that this person trusts you? and (c) If you were in trouble or were sad would you discuss it with this person? Response options for one item ranged from very well (1) to very badly (4). Response options for the other items ranged from all of the time (1) to never (5). Current parental monitoring was assessed with two items that asked about “the parent or guardian in the home who knows the most about [the youth]” (Cronbach’s α = .77): (a) Does this person know where you are when you are not at home? and (b) Does this person know who you are with when you are not at home? Current parental coercive discipline was assessed with two items that asked about “the parent or guardian in the home who knows the most about [the youth]” (Cronbach’s α = .61): (a) Does this person yell at you? and (b) Does this person take away your privileges? Response options ranged from all of the time (5) to never (1). We created a summation variable for each parent–child relationship characteristic to reflect more negative parent–child relationship characteristics (i.e., lack of emotional bond, lack of parental monitoring [possible range of 2-10], and greater coercive parenting [possible range of 2-10]). To make the items for emotional bond commensurate and, therefore, to allow for the creation of a composite variable, responses were standardized within items and then summed across items (range of −2.7 to 11.2).

Internet use and parental monitoring of Internet use

Youth reported the amount of time they use the Internet on a typical day using response options ranging from more than 3 hr (6) to 0 min (1). Youth also reported whether (0) or not (1) there was blocking or parental control software on the computer they use the most to access the Internet. Caregivers reported whether (0) or not (1) they currently have rules about how their child uses the Internet and how often they enforce these rules (almost always [1] to almost never [4]). These latter three variables were coded to reflect lack of Internet safety procedures.

Academic problems

Youth reported the average grades they earned in school during the past/current school year with larger values reflecting worse grades.

Demographic characteristics and process variables

Caregivers reported youths’ biological sex (female [1] vs. not [0]) and age. Youth reported their ethnicity (i.e., whether they are of Hispanic origin), race (i.e., whether they consider themselves White, etc.), whether they had responded honestly (not honest [1] vs. honest [0]), and whether they had been alone when completing the survey (not alone [1] vs. alone [0]).

Procedure

Wave 1 GuwM data were collected from 1,587 households between August and September, 2006. From these households, Wave 2 data were collected between November 2007 and January 2008 (n = 1,205) and Wave 3 data between August and November 2008 (n = 1,158). Herein we refer to the timing of the measurement of the data as Year 0, Year 1, and Year 2, respectively, to correspond with predicting the pattern of technology-based interpersonal victimization that occurred during the first (Year 1) and second year (Year 2) of a 2-year period based on characteristics occurring during Year 0, the year immediately prior to the 2-year period.

The sample was obtained from the Harris Poll Online (HPOL) opt-in panel (http://www.harrisinteractive.com/MethodsTools/DataCollection/HarrisPollOnlinePanel.aspx). Households were enrolled using a stratified random sample design based on youth age and sex. Youth were recruited through email contact with adult HPOL members and were required to be 10 to 15 years old, read English, have used the Internet in the last 6 months, and live in the household at least 50% of the time. Adult consent was obtained; adults completed a 5-min survey; and then youth assented and completed a 30-min survey. Youth received a US$15 gift certificate and caregivers US$10 at Years 0 and 1; and US$25 and US$15, respectively, at Year 2. Similar to widely cited national surveys (e.g., Pew Research Center, 2012), the response rate was 26%. The follow-up rate was 76% at Year 1 and 73% at Year 2. Additional methodological details are described in methodology reports (GuwM, 2007, 2008, 2009).

Using standardized, research-supported procedures (e.g., Schonlau et al., 2004), data were weighted statistically to reflect the population of adults with children ages 10 to 15 years old in the United States (Bureau of Labor Statistics & Bureau of the Census, 2006), and survey sampling weights adjusted for adults’ self-selection. Missing data and “refused” responses were imputed using a multiple imputation procedure (Rubin, 1987).¹ To reduce the likelihood of imputing truly nonresponsive answers, participants were required to have valid data (not “missing” or “don’t know”) for at least 80% of the survey questions asked of all youth. Consequently, five respondents were dropped from the Year 0 sample, nine from the Year 1 sample, and seven from the Year 2 sample.

The final analytic sample was constrained to the participants who provided valid data for Years 0, 1, and 2 (i.e., 1,018, 64% of the Year 0 respondents). As shown in Table 1, the weighted analytic sample was similar to the excluded sample in terms of demographic characteristics but differed on several psychosocial characteristics. The weighted and unweighted analytic samples differed only in proportion of Hispanic youth and average reported academic problems.

Statistical Methods

A series of ANOVAs examined the overall relationships between each Year 0 factor and pattern of technology-based interpersonal victimization over the 2-year period, while controlling statistically for survey process variables (i.e., being alone and honest while completing survey). We then identified latent risk factors underlying the individual Year 0 risk factors related to multi-item/measure constructs (e.g., experience with problem behavior) using principal components analysis with Varimax rotation, an eigenvalue of 1.0 as criterion for factor extraction, and a factor loading cutoff of |.40| for each item.

We then analyzed orthogonal factor scores of the identified latent risk factors (with greater factor scores reflecting more risk) and standardized scores of the individual risk factors (i.e., academic problems, female, and age) using profile analysis to examine how the risk factors—individual and latent—distinguish between the four technology-based interpersonal victimization groups while controlling statistically for survey process variables. Profile analysis, a special form of MANOVA, compares profiles of a group’s responses on a set of commensurate dependent measures. In particular, we are interested in the profile analysis tests of levels and parallelism. The test of levels examines whether groups differ in their overall score on the set of dependent measures, or, in this study, whether the average level of risk differed by victimization group. The test of parallelism examines whether groups differ in their profiles across the set of dependent variables, or, in this study, whether the shape of the profile in the level of risk on the different risk factors varies by victimization group. Here the multivariate F-test statistic is interpreted using degrees of freedom appropriate to this multivariate test. (See Tabachnick & Fidell, 2001, for a detailed discussion of this statistical procedure.)

Two additional multivariate regression analyses examined the relationships between Year 0 individual and latent risk factors and frequency of technology-based interpersonal victimization (a) during Year 1 and (b) during Year 2 while controlling statistically for survey process variables.

Results

More than half (56%) of youth reported technology-based interpersonal victimization at some point during Years 1 and 2, including 29% of youth who reported victimization during both years (i.e., re-victimized youth). The desisted victimized group was the smallest at 10% of youth; 17% of youth were later victimized. Furthermore, results of the individual ANOVAs, shown in Table 3, indicate that pattern of technology-based interpersonal victimization was associated with almost all the individual Year 0 predictors and had the strongest association with Year 0 technology-based interpersonal victimization as evidenced by the effect size (partial η²) of .14.²

Table 3.

Overall Relationships Between Year 0 Predictors and Pattern of Technology-Based Interpersonal Victimization Over Subsequent 2-Year Period.^a

	Not Victimized (n = 423)	Re-Victimized (n = 275)	Desisted Victimized (n = 101)	Later Victimized (n = 171)
Year 0 Predictor	M	M	M	M	Overall F	P	Partial η²
Female	0.44	0.61^b	0.49	0.51^b	6.66	<.001	.02
Age	12.21	13.12^b	12.73^b	12.52^b	16.57	<.001	.05
Current academic problems	6.38	6.21	6.37	6.15	1.69	.168	.01
Used alcohol in past year	0.07	0.20^b	0.13^b	0.07	10.33	<.001	.03
Used marijuana in past year	0.04	0.08	0.07	0.03	2.47	.061	.01
Current time spent on the Internet during typical day	3.03	4.04^b	3.66^b	3.23^b	34.92	<.001	.10
Currently do not have blocking software on the computer	0.56	0.63^b	0.62	0.50	2.67	.046	.01
Currently do not have household Internet rules	0.04	0.08^b	0.07^b	0.08^b	2.61	.050	.01
Current lack of parental enforcement of household Internet rules	1.47	1.69^b	1.80^b	1.59	4.75	.003	.02
Current lack of general parental monitoring	2.74	3.02^b	3.06^b	2.62	5.49	.001	.02
Current lack of youth emotional bond with caregiver	−0.57	0.96^b	0.43^b	−0.49	29.18	<.001	.08
Current coercive parental discipline	5.37	5.63^b	5.38	5.24	3.09	.027	.01
Delinquent behavior in past year	0.61	1.34^b	0.94^b	0.97^b	22.13	<.001	.06
Technology-based perpetration in past year	0.10	1.49^b	0.80^b	0.70^b	21.69	<.001	.06
Technology-based victimization in past year	0.31	3.24^b	2.36^b	1.14^b	52.74	<.001	.14
Traditional perpetration in past year	1.20	2.65^b	1.97^b	1.97^b	14.48	<.001	.04
Traditional victimization in past year	2.17	3.36^b	3.29^b	2.49	10.68	<.001	.03

Separate ANOVAs, one per Year 0 predictor, were used to examine the overall relationship between each predictor and technology-based interpersonal victimization group (i.e., not victimized, re-victimized, desisted victimized, and later victimized) while controlling for survey process variables.

This mean differs statistically significantly (p ≤ .05) from the mean for the not victimized group.

Results of the factor analysis, Table 4, indicated four factors reflecting experience with problem behavior (I), lack of Internet safety procedures (II), substance use (III), and negative parent–child relationship characteristics (IV). On average, across all the risk factors (individual and latent), the re-victimized group had the greatest average score (0.21) followed by the desisted victimized group (0.05), the later victimized group (−0.05), and, finally, the not victimized group (−0.16), F(3, 962) = 49.30, p< .001, partial η² = .13.

Table 4.

Varimax-Rotated Principal Components Factor Analysis of Year 0 Predictors to Identify Latent Risk Constructs.

Year 0 Predictor	I	II	III	IV
Delinquent behavior in past year	.66	−.02	.30	.20
Technology-based perpetration in past year	.78	.06	.22	.01
Technology-based victimization in past year	.79	.10	.10	.02
Traditional perpetration in past year	.86	.03	.14	.09
Traditional victimization in past year	.71	−.03	−.01	.14
Currently do not have blocking software on the computer	.02	.46	−.24	.10
Currently do not have household Internet rules	.02	.80	.15	−.15
Current lack of parental enforcement of household Internet rules	.01	.84	.13	−.04
Current time spent on the Internet during typical day	.10	.42	.01	.37
Used alcohol in past year	.27	.02	.79	.14
Used marijuana in past year	.21	.05	.85	.01
Current lack of general parental monitoring	.39	.07	.24	.40
Current lack of youth emotional bond with caregiver	.21	.12	.05	.79
Current coercive parental discipline	−.01	−.16	.02	.73
Eigenvalue	4.06	1.77	1.33	1.11
Percentage of variance	29.02	12.64	9.49	7.96

Note. Bold text indicates which Year 0 predictors loaded highly on each factor.

The profile, or pattern, of scores across the risk factors varied by technology-based interpersonal victimization group, multivariate F(18, 2877) = 3.76, p < .001, partial η² = .02. As shown in Figure 1, re-victimized youth had higher scores for being female, age, experience with problem behaviors, substance use, and negative parent–child relationships at Year 0 as compared with the other three groups. All four groups of youth varied the most from each other in terms of age, experience of problem behavior, and negativity of relationships with their parents at Year 0.

Figure 1.

Profiles of the four technology-based interpersonal victimization groups across individual and latent risk factors.

Results of the multivariate regression analyses examining predictors of amount of technology-based interpersonal victimization in the near (Year 1) and more distant (Year 2) future are shown in Table 5. These results indicate that being female, experience with problem behavior, substance use, and negative parent–child relationships at Year 0 were associated with greater frequency and/or type of technology-based interpersonal victimization during Years 1 and 2. However, the associations of technology-based interpersonal victimization with experience with problem behavior and substance use were stronger during Year 1. In addition, academic problems and lack of Internet safety procedures were associated with frequency of victimization during Year 2 only.³

Table 5.

Multivariate Regression Models^a Predicting Frequency of Technology-Based Interpersonal Victimization During Subsequent Years.

	Model 1: Frequency of Technology-Based Victimization During Year 1				Model 2: Frequency of Technology-Based Victimization During Year 2
	B	β	t	p	B	β	T	p
Intercept	2.65	—	13.38	<.001	2.96	—	14.22	<.001
Year 0 risk factor
Individual risk factor
Academic problems	0.03	.01	0.20	.842	0.41	.08	2.65	.008
Female	0.50	.10	3.54	<.001	0.57	.11	3.82	<.001
Age	0.19	.04	1.28	.201	0.15	.03	0.94	.346
Latent risk factor
Factor I: Experience with problem behavior	2.10	.32	10.75	<.001	1.59	.24	7.77	<.001
Factor II: Lack of Internet safety procedures	0.17	.04	1.21	.228	0.29	.06	2.00	.046
Factor III: Substance use	1.03	.21	7.09	<.001	0.57	.12	3.76	<.001
Factor IV: Negative parent–child relationship characteristics	0.85	.17	5.99	<.001	0.74	.15	4.99	<.001
Process variable
Not honest when responding to the survey	0.44	.02	0.57	.567	−0.02	<−.01	−0.02	.982
Not alone when responding to the survey	−0.70	−.07	−2.46	.014	−0.65	−.07	−2.17	.030

Note. Bold text indicates statistically significant (p ≤ .05) associations.

Analyses controlled for survey process variables.

Discussion

Findings highlight the importance of longitudinal data analyses to better understand youth experiences over time. Many factors reported in prior research to be related to concurrent technology-based interpersonal victimization are related to future technology-based interpersonal victimization as well. Of the individual risk factors examined, prior-year technology-based interpersonal victimization and Internet use have the strongest overall association with the pattern of technology-based interpersonal victimization during the next 2 years. Consequently, people working with youth should be particularly diligent in monitoring youth who use Internet excessively, and particularly responsive to incidences of technology-based interpersonal victimization to reduce the risk of future victimization.

Limiting or monitoring Internet use (which is relatively malleable) has the potential to be an immediately effective first response to incidences of technology-based interpersonal victimization. Whether Internet use is causally related to technology-based interpersonal victimization, or instead is related to it via a third variable, using this as an initial step could potentially disrupt further incidences in the near term until longer term solutions are identified and enacted. That said, some youth express reticence to talk to their parents about their online victimization experiences because they fear losing access to technology. Thus, action plans need to be thoughtful and tailored to the individual youth.

Results of the current study suggest that victimization is experienced chronically by youth in multiple ways. An important minority (one in four youth) is victimized over time. In addition, an important minority is victimized via traditional and technology-based means as evidenced by the strong interrelation between these types of victimization. This finding seems reminiscent of other recent research suggesting that because of their previous experiences (including prior victimization), some young people are more likely to be victims of multiple assaults over time (Finkelhor, Ormrod, & Turner, 2007). More attention needs to be paid to developing and testing effective intervention programs for youth who experience victimization chronically.

Findings also highlight that not all victims are completely blameless. Interpersonal victimization and perpetration are strongly interrelated as evidenced by their loading highly on one latent risk factor. Prevention efforts should focus on the reduction of peer aggression generally, as opposed to focused specifically on victims or perpetrators given that, oftentimes, these are the same youth.

The results of the current study indicate that the different types of victimized groups (e.g., re-victimized) were distinguished from each other in terms of overall score averaged across all the risk factors, as well as in the profiles, or patterns, of scores across the risk factors. Three risk factors in particular best distinguish the groups: age, experience of problem behaviors, and negative parent–child relationship. Although beyond the scope of the current article, these results could be used to direct future efforts to identify a relatively short assessment that accurately detects risk, for example, one that assesses age, experience of problem behaviors, and negative parent–child relationship. Furthermore, these risk factors, except for age, are malleable and could be reduced with intervention programs that, for example, help youth gain interpersonal skills and knowledge of how to translate them to technology-based and traditional environments, or inform parents about the importance of and how to engage in Internet safety procedures. Results also suggest that malleable risk factors are associated with the frequency of technology-based victimizations during the near and the more distant future. Some of these factors (e.g., Internet safety procedures) might be more malleable than others (e.g., negative parent–child relationships). However, each can be used to identify youth at risk for future victimization and the extent of that risk in terms of amount of victimization. In addition, each offers an opportunity to decrease that risk, for example, teaching Internet safety procedures in schools.

These findings should be considered within the study limitations. The analytic sample differed from the excluded respondents on a number of key characteristics. Thus, the pattern of re-victimization and the influence of the predictors might be different if all youth had been included. In addition, the response rate was 26%. HPOL data are consistently comparable with data that have been obtained from random telephone samples of the general population once propensity weighting and appropriate sample weights are applied (Schonlau et al., 2004), which suggests that findings would likely be similar if the sample were identified from a random digit dial methodology. In addition, a limited number of types of technology-based victimizations were measured and a few predictors had internal consistency below the recommended cutoff of .70. The influence of the predictors measured with greater internal consistency and of the predictors on other types of victimizations might be different. Finally, the data were self-reported online. We used multiple procedures to encourage honest and accurate reporting, however, to increase the confidence that the data accurately reflect the characteristics and experiences of the youth who participated in this study.

Conclusion

Given that more than half of youth in the current study report technology-based victimization during the 2-year period and more than one in four report repeated victimization, it is critical to identify easy and efficient ways of identifying youth at risk for technology-based victimization. These victimization experiences are related to risk factors that vary in their malleability and suggest opportunities for future intervention. Identifying these youth prior to the period of concern makes it possible to intervene with programs that effectively decrease the risk of future victimization.

Footnotes

Acknowledgements

We thank the entire Growing up With Media Study team: Internet Solutions for Kids, Harris Interactive, Johns Hopkins Bloomberg School of Public Health, and the Centers for Disease Control and Prevention (CDC). We also thank the families for their time and willingness to participate in this study.

Authors’ Note

The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention (CDC).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data collection for this manuscript was supported by Cooperative Agreement number U49/CE000206 from the Centers for Disease Control and Prevention (CDC) awarded to Michele L. Ybarra.

Notes

Author Biographies

Josephine D. Korchmaros, PhD, is the director of research methods and statistics at the University of Arizona’s Southwest Institute for Research on Women. She has conducted family, youth, social, health, marketing, and psychological research. Her areas of research include interpersonal relationships, sexual health, and addiction. She has played a key role in numerous federally funded research projects. Three of these are national surveys of adolescents examining the association between exposure to media violence and seriously violent behavior (CDC U49CE000206); the etiology of sexual violence in adolescence (R01 CE001543); and how technology plays a protective as well as risk-enhancing role in the lives of lesbian, gay, bisexual, and transgender (LGBT) adolescents (R01 HD057191). The others include research projects focused on the implementation and/or evaluation of health-related intervention programs including, for example, a national, cross-site evaluation of juvenile drug courts (LCFRD11C0007), an Internet-based HIV prevention program for adolescents living in Mbarara, Uganda (R01 MH80662), a recovery-oriented support system for adolescent girls in substance abuse treatment (H79 T1018406), a text-messaging-based smoking cessation program among adults living in Ankara, Turkey (R01 TW007918), and an HIV and substance use relapse prevention program for adolescents (H79 TI20138).

Kimberly J. Mitchell, PhD, is a research associate professor of psychology at the Crimes against Children Research Center (CCRC), located at the University of New Hampshire. Her areas of research focus on youth Internet victimization and how technology can be used as a tool for prevention and intervention. She has been studying Internet use among youth for more than 12 years. She has directed and/or codirected several projects including the First and Second Youth Internet Safety Studies; the Survey of Internet Mental Health Issues; the First, Second, and Third National Juvenile Online Victimization Studies; and the National Juvenile Prostitution Study. She was the principal investigator (PI) of the Third Youth Internet Safety Survey (funded by Office of Juvenile Justice and Delinquency Prevention [OJJDP]), a grant to investigate the commercial exploitation of children through the Internet and a gift from Microsoft to explore the role of technology in human trafficking. She is the author of more than 60 peer-reviewed papers in her field and has spoken at numerous national conferences.

Michele L. Ybarra, MPH PhD, is the president and research director of Center for Innovative Public Health Research (formerly known as Internet Solutions for Kids), a nonprofit research organization in the United States centered on understanding the impact on and opportunities for adolescent health represented by new technologies. She is internationally recognized for her work in the field of Internet victimization, including cyberbullying and Internet harassment and unwanted sexual solicitation. She is the PI for Growing Up With Media, a national, longitudinal survey of adolescents in the United States to identify the associations between violence in new media and seriously violent behavior (CDC U49CE000206). This study has recently been extended to examine characteristics associated with the manifestation of sexually violent behavior over time (R01 CE001543). She also is leading a study of the benefits as well as risks posed by the Internet for LGBT youth (R01 HD057191). She is also the PI for the CyberSenga project in Mbarara, Uganda, a novel, Internet-based HIV prevention program for youth (R01 MH80662). In addition, she is the PI of an National Institutes of Health (NIH)/Fogerty Foundation study of SMS Turkey, a text messaging-based smoking cessation program for adult smokers living in Ankara, Turkey (R01 TW007918), and a sister project in the United States (R21 CA135669).

References

Bureau of Labor Statistics & Bureau of the Census. (2006, July 5). Current Population Survey. Retrieved from http://www.bls.gov/cps/demographics.htm

Centers for Disease Control and Prevention. (2005). Measuring violence-related attitudes, behaviors, and influences among youths: A compendium of assessment tools (2nd ed.). Atlanta, GA: Author.

Centers for Disease Control and Prevention. (2006). Youth Risk Behavior Surveillance—United States. Morbidity and Mortality Weekly Report, 55, 1-108.

Cohen

L. E.

Felson

(1979). Social change and crime rate trends: A routine activity approach. American Sociological Review, 44, 588-608.

Finkelhor

Hashima

(2001). The victimization of children and youth: A comprehensive overview. In White

(Ed.), Handbook on youth and justice (pp. 49-78). New York, NY: Kluwer Academic Publishers.

Finkelhor

Mitchell

K. J.

Wolak

(2000). Online victimization: A report on the nation’s youth (No. 6-00-020). Alexandria, VA: National Center for Missing & Exploited Children.

Finkelhor

Ormrod

R. K.

Turner

H. A.

(2007). Poly-victimization: A neglected component in child victimization. Child Abuse & Neglect, 31, 7-26.

Growing up With Media: Wave 1 Online Methodology Report. (2007, March). Retrieved from http://innovativepublichealth.org/wp-content/uploads/2013/04/Wave-1-Methodology-Report.pdf

Growing up With Media: Wave 2 Online Methodology Report. (2008, April). Retrieved from http://innovativepublichealth.org/wp-content/uploads/2013/06/Wave-2-Methodology-Report.pdf

10.

Growing up With Media: Wave 3 Online Methodology Report. (2009, July). Retrieved from http://innovativepublichealth.org/wp-content/uploads/2013/06/Wave-3-Methodology-Report.pdf

11.

Hamby

Finkelhor

Turner

Kracke

(2011). The Juvenile Victimization Questionnaire toolkit. Retrieved from http://cola.unh.edu/ccrc/juvenile-victimization-questionnaire

12.

Jones

L. M.

Mitchell

K. J.

Finkelhor

(2012). Trends in youth internet victimization: Findings from three youth internet safety surveys 2000-2010. Journal of Adolescent Health, 50, 179-186.

13.

Lenhart

Madden

Macgill

Smith

(2007, December). Teens and social media: The use of social media gains a greater foothold in teen life as they embrace the conversational nature of interactive online media. Washington, DC: Pew Internet & American Life.

14.

Mitchell

K. J.

Finkelhor

Wolak

(2007). Youth internet users at risk for the most serious online sexual solicitations. American Journal of Preventive Medicine, 32, 532-537.

15.

Pew Research Center for the People & the Press. (2012). Assessing the representativeness of public opinion surveys. Washington, DC: Author. Retrieved from http://www.people-press.org/2012/05/15/assessing-the-representativeness-of-public-opinion-surveys/

16.

Rubin

D. B.

(1987). Multiple imputation for nonresponse in surveys. New York, NY: J. Wiley & Sons

17.

Schonlau

Zapert

Simon

L. P.

Sanstad

K. H.

Marcus

S. M.

Adams

(2004). A comparison between response from a propensity-weighted web survey and an identical RDD survey. Social Science Computer Review, 22, 128-138.

18.

Spatz Widom

Czaja

S. J.

Dutton

M. A.

(2008). Childhood victimization and lifetime revictimization. Child Abuse & Neglect, 32, 785-796.

19.

Tabachnick

B. G.

Fidell

L. S.

(2001). Using multivariate statistics (4th ed.). Needham Heights, MA: Allyn and Bacon.

20.

Turner

H. A.

Finkelhor

Ormrod

(2006). The effect of lifetime victimization on the mental health of children and adolescents. Social Science & Medicine, 62, 13-27.

21.

von Hippel

P. T.

(2007). Regression with missing Ys: An improved strategy for analyzing multiply imputed data. Sociological Methodology, 37, 83-117.

22.

Williams

K. R.

Guerra

N. G.

(2007). Prevalence and predictors of internet bullying. The Journal of Adolescent Health, 41(6 Suppl. 1), S14-S21.

23.

Ybarra

M. L.

Diener-West

Leaf

P. J.

(2007). Examining the overlap in internet harassment and school bullying: Implications for school intervention. Journal of Adolescent Health, 41(6 Suppl. 1), S42-S50.

24.

Ybarra

M. L.

Mitchell

K. J.

(2004). Online aggressor/targets, aggressors, and targets: A comparison of associated youth characteristics. Journal of Child Psychology and Psychiatry, 45, 1308-1316.

25.

Ybarra

M. L.

Mitchell

K. J.

(2007). Prevalence and frequency of internet harassment instigation: Implications for adolescent health. Journal of Adolescent Health, 41, 189-195.

26.

Ybarra

M. L.

Mitchell

K. J.

Korchmaros

J. D.

(2011). National trends in exposure to and experiences of violence on the internet among children. Pediatrics, 128, e1376-e1386.

27.

Ybarra

M. L.

Mitchell

K. J.

Wolak

Finkelhor

(2006). Examining characteristics and associated distress related to internet harassment: Findings from the second youth internet safety survey. Pediatrics, 118, 1169-1177.