Family functioning patterns predict teenage girls’ sexting

Definition and Prevalence of Sexting

Sexting refers to the exchange of sexually suggestive or provocative text messages, images, and videos via Internet or smartphone (Barrense-Dias, Berchtold, Suris, & Akre, 2017; Chalfen, 2009). The prevalence of sexting among adolescents is increasing worldwide. Studies published prior to 2013 found that 20% of adolescents and 33% of young adults engaged in sexting (Eurispes & Telefono Azzurro, 2012; National Campaign to Prevent Teen and Unplanned Pregnancy & CosmoGirl.com, 2008). These percentages are now between 60 and 80 (Drouin, Coupe, & Temple, 2017; Morelli, Bianchi, Baiocco, Pezzuti, & Chirumbolo, 2016a).

Different kinds of sexting have been described. Calvert (2009) introduced a distinction between primary sexting, or consensual sharing of sexual content related to oneself, and secondary sexting, or forwarding of sexts received from others, who may be unaware and may not have consented to the sharing of sexual material. Wolak, Finkelhor, and Mitchell (2012) distinguished between experimental and aggravated sexting. Experimental sexting, which is predominantly consensual for the senders, is driven by developmental needs related to socialization and exploration of sexuality and identity (Levine, 2013). Engaging in sexting with a partner in the context of a trusting, romantic relationship is probably the most frequent sexting behavior among young people. It is commonly conceived as a way to experiment with sexuality and intimacy in the relationship (Döring, 2014). Aggravated sexting, in contrast, implies deviant and aggressive intentions (such as sexting that is coerced; Drouin & Tobin, 2014). Morelli, Bianchi, Baiocco, Pezzuti, and Chirumbolo (2016b) investigated a specific dimension of aggravated secondary sexting consisting of forwarding sexts from someone without his or her consent, which we refer to as nonconsensual forwarding of sexts. This behavior may be a cyber form of sexual harassment (Machimbarrena et al., 2018) and can include dating aggression and revenge (Bates, 2017; Bianchi, Morelli, Nappa, Baiocco, & Chirumbolo, 2018; Morelli, Bianchi, Cattelino, et al., 2017) and cyberbullying (Medrano, Lopez Rosales, & Gámez-Guadix, 2018).

Importantly, sexting behaviors increase with age (Dake, Prize, Maziarz, & Ward, 2012; Strassberg, McKinnon, Sustaita, & Rullo, 2013). Specifically, both sending and receiving sexts increase from early (13–16 years old) to late adolescence (17–20 years old; Morelli, Bianchi, Baiocco, Pezzuti, & Chirumbolo, 2017), and sharing of own sexts with others is more likely to occur in young adulthood than adolescence (21–30 years old; Bianchi, Morelli, Baiocco, & Chirumbolo, 2016).

Positive and Risky Facets of Sexting

The literature on sexting has been divided between a “normalcy” and a “deviance discourse” (Döring, 2014). Within the normalcy discourse, Levine (2013) suggested that sexting can address developmental needs during the transition from adolescence to young adulthood. Self-presentation on social media facilitates comparisons with peers and with media models (Valkenburg, Peter, & Schouten, 2006), supporting the exploration of sexuality and identity construction. Similarly, sexting can be used by young people to solicit confirmation of body adequacy and increase self-esteem (Bianchi, Morelli, Baiocco, & Chirumbolo, 2017). Some vulnerable groups, such as sexual minorities, reported using sexting and Web-based communication to meet potential partners and friends (Chong, Zhang, Mak, & Pang, 2015), and thus might be considered protective against social stigmatization and prejudice (Meyer, 2003). Moreover, sexting among romantic partners was found to increase passion and intimacy (Drouin & Tobin, 2014), especially in long-distance relationships (Renfrow & Rollo, 2014). Finally, some studies have reported that sexters and non-sexters do not differ with respect to psychological well-being (e.g., anxiety, depression; Morelli et al., 2016a).

The deviance discourse, in contrast, regards sexting as a risky behavior due to the potential negative impact on psychological well-being, especially during adolescence (Klettke, Hallford, & Mellor, 2014). Sexting has been linked to alcohol and substance abuse, sexual risk-taking, and impulsivity-related traits, such as sensation seeking and negative urgency (the tendency to react impulsively to negative emotions; Dir, Koo, & Cyders, 2013). Negative urgency increases during adolescence, leading to risk-taking behaviors and difficulties in emotional regulation (Cyders & Smith, 2008). In line with these findings, sexting has been related to emotional problems (Ševčíková, 2016), emotional dysregulation, and low emotional self-efficacy (Trub & Starks, 2017). Yoder, Hansen, and Precht (2018) shed light on emotional motivations for sexting, specifically anger, loneliness, and attention seeking. These sexting motives, predicted by negative family experiences during childhood, were correlated to sexting outside of a dating relationship and with dating violence. Overall, these results suggest that sexting for emotion regulation is a risky behavior.

In addition to sexting for emotion regulation, the sharing of sexts should be considered a risky behavior, because it increases the probability of uncontrolled online diffusion of sexual material, which can, in turn, lead to cyber harassment and cyberbullying (Bryce & Klang, 2009). The link between sexting and cybervictimization is well-documented, and the unwanted dissemination of sexts can lead to severe psychological distress, including suicide attempts (Dake et al., 2012). Furthermore, Temple et al. (2012) found that, among teenage girls, sexting was associated with having multiple sexual partners and with other sexually risky behaviors.

Sexting in Teenage Girls

Boys more frequently request and receive sexts than do girls, whereas girls are more likely to send sexts than are boys (Ybarra & Mitchell, 2014). Furthermore, girls are more likely than boys to send sexts under pressure from peers or partners, to prove their commitment or to avoid the breakup of a relationship (Van Ouytsel, Van Gool, Walrave, Ponnet, & Peeters, 2017). Moreover, girls more frequently sext to gain a partner’s attention (Van Ouytsel et al., 2017), whereas boys are more likely to forward sexts they have received to other people (Garcia et al., 2016).

Sexting may be riskier for girls than for boys due to cultural pressure, gendered sexual expectations, and sexual double standards (Ringrose, Harvey, Gill, & Livingstone, 2013). According to Ringrose, Harvey, Gill, and Livingstone (2013), sexting is interpreted differently in boys and girls; sexting is regarded as a positive expression of one’s gender roles when performed by boys, but not girls. Sexual double standards can be directly connected to the sexual objectification theory (Fredrickson & Roberts, 1997), which accounts for the pressure that cultural and media models place on women and girls to conform to unrealistic and sexualized physical standards. According to the American Psychological Association (2007), girls are exposed to sociocultural standards of sexual attractiveness and begin to internalize them. Eventually, many girls view their body an object of others’ sexual desire, which is associated with dysfunctional behavioral and psychological consequences, both in the short (McKenney & Bigler, 2016) and long term (Impett, Henson, Breines, Schooler, & Tolman, 2011). In particular, young girls are more vulnerable to sexual harassment (Zurbriggen & Roberts, 2013). A study of Italian 13- to 20-year-olds demonstrated that sexual self-objectification predicted some types of sexting (Bianchi et al., 2017). In recent decades, Italy has been characterized by the pervasive diffusion of sexualized models of women’s bodies via television and other media (Benini, 2012).

Family Context, Sexual Risk-Taking, and Sexting

A literature review on parenting and adolescent sexual behavior (Kincaid, Jones, Sterrett, & McKee, 2012) examined the negative effects of overparenting and psychological and behavioral control on sexual risk-taking during adolescence. Intrusive and manipulative parental control inhibits the development of decision-making skills, increasing vulnerability to sexual risky behaviors (Cummings, Davies, & Campbell, 2000). Furthermore, overcontrolling parenting can exacerbate problem behaviors and relational difficulties of adolescents (Pace, D’Urso, & Zappulla, 2018a, 2018b). In contrast, positive family climates, characterized by warmth, support, and good parent–child communication, predict positive adolescent outcomes (Baumrind, 1995). As suggested by Longmore, Manning, and Giordano (2001), positive family contexts provide an adaptive model for intimate relationships and help adolescents to develop their self-regulation abilities, interpersonal security, and respect for interpersonal boundaries in intimate relationships. Specifically, parental support, care, and closeness during adolescence were found to improve psychological well-being, self-esteem, and positive self-concept in adulthood, leading to well-adjusted adult behavior (Baumrind, 1995; Roberts & Bengtson, 1996).

Previous studies suggest that parental monitoring, family communication, and parenting styles are related to sexual risk-taking during adolescence (Huebner & Howell, 2003; Yang et al., 2007). In particular, good parent–child communication, balanced warmth, and monitoring were found to protect against sexual risk-taking during adolescence (Diiorio, Pluhar, & Belcher, 2003). According to Yang et al. (2007), communication with parents can shape adolescents’ risk-taking, whereas perceived communication difficulties increase early sexual activity in girls. Several studies have shown that the influence of family climate operates differently in male and female adolescents, with good communication, good relationships, and closeness being specific protective factors for girls (Buhi & Goodson, 2007). According to Diiorio, Pluhar, and Belcher (2003), the quality and the content of parent–child communication underpin the development of sexual decision-making during adolescence, potentially promoting healthy sexuality and reducing risky behaviors. Family communication about sexuality plays a key role in the development of healthy sexuality (Pluhar & Kuriloff, 2004). Parental availability and good quality of family relationship can also protect adolescents against problematic and risky online behaviors (Trumello, Babore, Candelori, Morelli, & Bianchi, 2018; Xiuqin et al., 2010).

We could locate only two studies of the influence of parental monitoring and family rules on sexting in adolescents (12–18 years old): a study by Campbell and Park (2014) with a U.S. sample and a study by West et al. (2014) with a Peruvian sample. Strict parental control was found to be ineffective in reducing sexting (Campbell & Park, 2014) and in avoiding dysfunctional Internet use (Trumello et al., 2018). However, sharing clear rules about sexting in family was found to protect adolescents against sexting behaviors (West et al., 2014). Nevertheless, developmental research has not yet investigated the role of specific family dynamics in adolescent sexting.

Hence, the aim of this study was to investigate the association between family functioning and teenage sexting, focusing on girls as an at-risk group (Ringrose et al., 2013), within the theoretical framework of the circumplex model of marital and family systems (Olson, Sprenkle, & Russell, 1979). This model conceptualizes family functioning in two main dimensions: cohesion and flexibility. Cohesion is defined as the level of emotional bonding among family members and the amount of individual autonomy they have. Flexibility is defined as the extent to which the family adapts rules and roles to deal with developmental and situational stressors. The model was subsequently extended to include four additional dimensions of dysfunctional family systems: rigidity, chaos, enmeshment, and disengagement. In rigid families, a leader makes all decisions, roles and rules are unchangeable, and negotiation is impossible. In chaotic families, roles are indistinct, rules unclear, and decisions tend to be impulsive. In enmeshed families, there is an excessive emotional closeness, individual space is limited, and members are focused on intra-family relationships. Finally, members of disengaged families show extreme emotional distance from one another and are largely independent with a lack of mutual support. Family communication has also been added to the model, as an important variable of family functioning (Olson, 2004; Olson & Gorall, 2006).

Aims and Hypotheses

Based on the literature reviewed above, and using the circumplex model of marital and family systems (Olson et al., 1979), we hypothesized that specific patterns of family functioning are related to specific sexting behaviors. Specifically, we investigated the possible differences in family functioning between girls who engaged in sexting and girls who did not, hypothesizing that a positive family climate (characterized by high levels of cohesion, flexibility, and good communication, and by low levels of rigidity, enmeshment, chaos, and disengagement) would protect against teenage sexting. Furthermore, we expected that a positive family climate would be related to experimental sexting (i.e., sharing sexts with a dating partner), in accordance with studies on the role of family in providing models for adaptive romantic relationships characterized by intimacy and communication (Feldman, Gowen, & Fisher, 1998). Conversely, dysfunctional family contexts (characterized by low levels of cohesion, flexibility, and communication and high levels of rigidity, enmeshment, chaos, and disengagement) were expected to be related to risky sexting (i.e., sharing sexts with a large number of people; sexting for emotion regulation) and aggravated sexting (i.e., nonconsensual forwarding of sexts). Specifically, we hypothesized that the main predictor of aggravated sexting would be family enmeshment, which might lead girls to fail to respect privacy limits.

Participants

Participants were 250 teenage girls (M_age = 16.36 years; SD_age = 1.88; age range = 13–20 years) recruited from Italian public high schools. Most participants were Italian (90%) and the remaining 10% were ethnic minorities. The mean parental education level was medium-high, with 40% of mothers and 38% of fathers having a high school diploma and the 44% of mothers and 39% of fathers having a university degree. Most participants (82%, n = 206) reported being exclusively heterosexual. Over one third (38%, n = 96) were currently involved in a dating relationship and another 42% (n = 105) had been in the past; the remaining 20% (n = 49) had never had a romantic partner. After participants (and their parents for those under 18 years of age) provided informed consent, participants completed an online survey at their school, which afforded respondents privacy and allowed for supervision (Clark Newman et al., 2002).

Measures

Overview of Data Analyses

Cronbach’s αs for study variables were computed. All were acceptable and appear in Table 1. Participants were divided in two age groups—early adolescents (13–16 years; n = 138; M_age = 14.9, SD_age = 0.86) and late adolescents (17–20 years; n = 112; M_age = 18.1; SD_age = 1.11)—to investigate age differences on all study variables, with multivariate analysis of variance (MANOVA) and χ² analyses. Correlations between variables were calculated. Because they are not central to our hypothesis testing, these appear in online supplemental materials. Non-normal variables (number of people girls share sexts with; nonconsensual forwarding of sexts; sexting for emotion regulation) were logarithmically transformed before testing hypotheses about regression relations. Finally, a set of regressions was used to determine which family functioning variables (i.e., communication, flexibility, cohesion, disengagement, chaos, enmeshment, and rigidity) predicted the various kinds of sexting, controlling for the effect of age (continuous score). We confirmed that our data met the assumptions of hierarchical multiple regression and the variance inflation factor scores ranged from 1.05 to 3.15, showing no problems with multicollinearity. Logistic regression was used for dichotomous dependent variables and linear regression for continuous dependent variables.

OPEN IN VIEWER

Table 1.

Means and Standard Deviations by Age Groups.

	Cronbach’s α	Early adolescents, M (SD)	Late adolescents, M (SD)
Family variables
Communication	.82	3.78 (0.90)	3.51 (0.95)
Cohesion	.78	3.97 (085)	3.93 (0.84)
Flexibility	.60	3.56 (0.79)	3.47 (0.77)
Disengagement	.67	1.91 (0.74)	1.89 (0.81)
Chaos	.78	2.22 (0.96)	2.23 (1.02)
Enmeshment	.62	2.43 (0.98)	2.62 (0.98)
Rigidity	.73	2.81 (0.88)	2.46 (0.77)
Sexting behaviors
Number of people sexting	—	0.31 (0.44)	0.55 (0.48)
Nonconsensual forwarding of sexts	.67	1.82 (0.10)	1.80 (0.04)
Sexting for emotion regulation	.88	1.25 (0.32)	1.36 (0.40)
N		138	112

Note. Item mean scores were computed for each variable. Item options for family variables—communication, cohesion, flexibility, disengagement, chaos, enmeshment, and rigidity—were as follows: 1 (strongly disagree), 2 (generally disagree), 3 (undecided), 4 (generally agree), and 5 (strongly agree). Item options for the number of people sexting—the number of people with whom girls have shared sexts—were as follows: 0 (no one), 1 (1 person), 2 (2–3 people), 3 (4–5 people), and 4 (more than 5 people). Item options for nonconsensual forwarding of sexts and for sexting for emotion regulation were as follows: 1 (never), 2 (seldom), 3 (sometimes or monthly), 4 (often or weekly), and 5 (frequently or daily).

Frequencies

Overall, 44% of participants (n = 109) reported to have engaged in sexting at least once. About a quarter of the sample (24%, n = 61) had only shared sexts with one person, 15% (n = 38) had shared sexts with two people, 4% (n = 9) with three to five people, and only one girl (less than 1%) reported having shared sexts with more than five people. Twenty participants (8%) reported nonconsensual forwarding of sexts at least once and 30% (n = 76) reported that they had sexted for emotion regulation at least once. Finally, of those who were or had been involved in a dating relationship, 37% (n = 75) had shared sexts with a dating partner at least once.

Age Differences

A MANOVA investigating age differences in family functioning variables yielded a multivariate effect, Wilks’s λ = .91, F(7, 242) = 3.28, p = .002, η² = .09, indicating an overall effect of age on the dependent variables. Early adolescents had higher mean scores than late adolescents for family communication, F(1, 249) = 5.41, p = .02, η² = .02, and rigidity, F(1, 249) = 10.91, p = .001, η² = .04. The second MANOVA on three sexting behaviors (number of people girls share sexts with; nonconsensual forwarding of sexts; sexting for emotion regulation) yielded a multivariate effect of age on the dependent variables, Wilks’s λ = .91, F(3, 246) = 8.26, p < .001, η² = .09. Late adolescents shared sexts with more people than early adolescents, F(1, 249) = 17.21, p < .001, η² = .06, and were more likely to sext for emotion regulation, F(1, 249) = 5.27, p = .02, η² = .02. Late adolescents were also more likely to have engaged in sexting than early adolescents (59% vs. 31%), χ²(1) = 19.39, p < .001. Finally, in the subsample of girls who had experience of a dating relationship, late adolescents were more likely to have shared sexts with a partner than early adolescents (48% vs. 27%), χ²(1) = 9.72, p = .002 (see Table 1).

Family Patterns in Sexting Behaviors

The first logistic regression investigated which family functioning variables predict engaging in sexting, controlling for age effects. The model accounted for 23.2% of the variance, Nagelkerke’s R² = .23, χ²(8) = 47.40, p <.001. Engaging in sexting was positively predicted by age, b = .41, p < .001, and negatively predicted by communication, b = −.56, p = .02. Using the subsample of girls who reported having (or having had) a dating partner (n = 201), a logistic regression analysis investigated which family functioning variables predict the sharing of sexts with partner (vs. not sharing), controlling for age effects. The results showed that sexting with a partner was positively predicted by age, b = .42, p < .001, and flexibility, b = .77, p = .008, which accounted for 23.1% of the variance, Nagelkerke’s R² = .23, χ²(8) = 37.25, p < .001 (see Table 2).

OPEN IN VIEWER

Table 2.

Family Functioning Patterns Related to the Engaging in Sexting and Sexting With Partner.

Predictors	Sexting behaviors
	Engaging in sexting				Sexting with partner
	B	SE	OR	95% CI	B	SE	OR	95% CI
Age	.41***	.08	1.50	[1.28, 1.76]	.42***	.09	1.53	[1.27, 1.84]
Communication	−.56*	.24	0.57	[0.36, 0.91]	−.29	.26	0.75	[0.45, 1.25]
Cohesion	.05	.30	1.05	[0.59, 1.88]	−.38	.33	0.68	[0.36, 1.30]
Flexibility	.11	.24	1.12	[0.70, 1.78]	.77**	.29	2.17	[1.23, 3.84]
Disengagement	.11	.29	1.12	[0.64, 1.95]	.32	.33	1.38	[0.72, 2.65]
Chaos	−.14	.19	0.87	[0.60, 1.26]	−.08	.22	0.93	[0.61, 1.42]
Enmeshment	.14	.15	1.15	[0.86, 1.53]	.11	.17	1.12	[0.81, 1.55]
Rigidity	−.10	.17	0.91	[0.65, 1.27]	.10	.20	1.10	[0.74, 1.64]
N	250				201

Note. OR = odds ratio; CI = 95% confidence interval. Logistic regression coefficients are reported.

*p < .05. **p < .01. ***p < .001.

Three further linear regressions were run to investigate which family functioning variables predicted (a) the number of people with whom girls share sexts, (b) nonconsensual forwarding of sexts, and (c) sexting for emotion regulation, controlling for age effects. The number of people with whom girls share sexts was predicted by age, β = .30, p < .001, and family communication, β = −.22, p = .01, which together accounted for 19.6% of the variance, R = .44, p = .005. The unique predictor of nonconsensual forwarding of sexts was enmeshment, β = .20, p = .002, which accounted for 7.1% of the variance, R = .27, p = .01. Finally, sexting for emotion regulation was predicted by age, β = .17, p = .01, and family communication, β = −.21, p = .02, which accounted for 10.8% of the variance, R = .33, p = .01 (see Table 3).

OPEN IN VIEWER

Table 3.

Family Functioning Patterns Related to Sexting Behaviors.

Predictors	Sexting behaviors
	Number of people sexting		Nonconsensual forwarding of sexts		Sexting for emotion regulation
	β	95% CI	β	95% CI	β	95% CI
Age	.30***	[.05, .11]	−.08	[−.01, .002]	.17**	[.01, .06]
Communication	−.22*	[−.20, −.02]	−.09	[−.02, .01]	−.21*	[−.15, −.01]
Cohesion	−.04	[−.14, .09]	−.01	[−.02, .02]	−.01	[−.09, .09]
Flexibility	.11	[−.20, −.02]	.14	[−.002, .03]	.11	[−.02, .12]
Disengagement	−.002	[−.02, .16]	.16	[−.003, .03]	.12	[−.03, .14]
Chaos	.05	[−.04, .10]	−.05	[−.02, .008]	−.01	[−.06, .05]
Enmeshment	.09	[−.01, .10]	.20**	[.01, .02]	.09	[−.01, .08]
Rigidity	−.09	[−.12, .01]	.03	[−.01, .01]	.02	[−.04, .06]
R 2	.20**		.07**		.11**

Note. CI = 95% confidence interval. Standardized regression coefficients are reported. N = 250.

*p < .05. **p < .01. ***p < .001.