A Longitudinal Investigation of Observed Adolescent Text-Based Sexting and Adjustment

Abstract

Despite strong concerns that sexting poses risks for adolescents’ well-being, previous research finds mixed results. Moreover, these studies rely heavily on self-report measures and cross-sectional designs. This study utilizes observational methods to examine longitudinal relations between text-based sexting and both negative and positive indicators of psychosocial adjustment. An ethnically diverse sample of 197 adolescents was provided smartphones that captured their text messages across high school, from 2008–2012. Two, two-day samples of text messages from grades 9–12 were content-coded (468,201 total observations). Sexting was defined as sending and receiving text-based statements about past, present, or hypothetical sexual behaviors occurring between dyadic texting partners. Each year, adolescents reported on negative (internalizing, externalizing, and social problems; borderline personality features) and positive (life satisfaction, group belongingness, and positive self-perceptions) indicators of their psychosocial adjustment. Using concurrent and longitudinal multilevel models, greater sexting at one time point than one’s average (within-person) was only associated with lower levels of group belongingness within the same year. For girls only, text-based sexting more than others on average (between-person) was related to increased externalizing symptoms and borderline personality disorder features, as well as decreased life satisfaction, group belongingness, self-perceived social competence, and global self-worth. No within-person differences by gender emerged, nor did longitudinal associations. Future studies and intervention efforts should examine when and why higher sexting might be related to psychosocial problems among girls.

Contemporary adolescents are increasingly living their lives on digital devices. Ninety-five percent of US teens own or have access to a smartphone and 45% of teens report being online almost constantly (Pew Research Center, 2018). Text messaging is a critical component of adolescents’ digital communication; 35% of US teens report that texting is their favorite way to communicate with friends, even above in-person communication (Common Sense Media, 2018). Co-construction theory posits that adolescents may use digital communication to explore key developmental processes (Subrahmanyam et al., 2006). Sexual exploration is a developmental process for many adolescents (Kroger, 2006), given pubertal onset, increased interest in and contact with romantic partners, and peer romantic socialization (Tolman & McClelland, 2011). One way in which adolescents may co-construct their online and offline sexual exploration is via sexting, or communicating about sex via text messaging.

With approximately one quarter of US adolescents engaging in sexting (Madigan et al, 2018), parents and policy makers alike fear the potential harms for youth (Best & Bogle, 2014), Yet, research evidence is mixed. On one hand, sexting may be associated with poorer mental health, humiliation, bullying, harassment, or non-consensual sharing/sending of sexts (e.g., Dake et al., 2012; Krieger, 2017). On the other hand, sexting may be perceived as normative (Symons et al., 2018) and provide a physically safer avenue for sexual exploration as compared to in-person sexual behavior (Campbell & Park, 2014; Holmes et al., 2020; Lippman & Campbell, 2014). To investigate sexting from both problem-focused and normative development-focused perspectives, this study employed longitudinal, observational methods to explore adolescents’ sexting behaviors in relation to both positive and negative indicators of psychosocial adjustment related to internal psychological health, external behaviors, and social wellbeing.

Defining and Measuring Sexting

Our current understanding of sexting is limited by the inconsistent operational definitions and methods used in previous research (Barrense-Dias et al., 2017; Mori et al., 2019). Sexting is often defined as sending nude- or nearly-nude pictures or videos (Barrense-Dias et al., 2017; Van Ouytsel et al., 2019). Sexting may also be defined more broadly to encompass word-based messages about sexual behaviors that partners intend to engage in or have previously engaged in (Champion & Pedersen, 2015; Mori et al., 2019). We measure sexting in line with this latter definition, in that sexting is operationalized as text-based messages about sexual behaviors that occurred within-dyad (i.e., involving the sender and receiver rather than anyone outside the conversation).

Considering all definitions of sexting together, meta-analyses suggest that, among adolescents, the prevalence of sending a sext is 14% and receiving a sext is 30% (Madigan et al., 2018; Molla-Esparza et al., 2020). Narrowing to the few text-based sexting studies, meta-analytic evidence suggests that 22% of adolescents send and 37% receive text-based sexts (Molla-Esparza et al., 2020). Estimates may vary by age, with older adolescents sexting more than younger adolescents (Baumgartner et al., 2014; Choi et al., 2019; Dake et al, 2012; Molla-Esparza et al., 2020). Self-reports of sexting may be biased by social desirability. Objective and observed sexting measures are less prone to bias; however, it is not ethical to directly observe most of adolescents’ image-based sexting given that it most often depicts sexual content produced by minors, which is legally considered child pornography (Bradley et al., 2020; Judge, 2012).

Though related, text-based sexting may be less risky than image-based sexting as it is often less explicit (Champion & Pedersen, 2015) and as image-based sexting may confer unique social and legal risks for adolescents (e.g., revenge pornography, nonconsensual sharing; Krieger, 2017). Given the lower risk level, text-based sexting may be related to more normative sexual exploration and associated with fewer negative adjustment outcomes than image-based sexting. Text-based sexting may even be advantageous if it functions similarly to offline sexual communication, which has been found to relate to greater sexual self-efficacy and safer sex practices (for a meta-review, see Noar et al., 2006). Yet, if text-based sexting is even considered, it is almost always confounded and combined with image-based sexting (for exceptions, see Brinkley et al., 2017; West et al., 2014). This study examines text-based dyadic sexting to explore if this potentially more frequent type of sexting relates to adolescents’ adjustment.

Sexting and Psychosocial Adjustment

Two theoretical perspectives can inform how text-based sexting may relate to adolescent adjustment. First, co-construction theory posits that relations between sexting and adjustment should parallel relations between offline sexual behavior and adjustment (positive or negative) and that behaviors that occur online should reciprocally relate to offline behaviors or experiences (Subrahmanyam et al., 2006). Second, problem behavior theory postulates that different problematic behaviors stem from the same underlying cause, and that individuals who are experiencing one problem behavior are more likely to exhibit other problems (Jessor & Jessor, 1977). Thus, if text-based sexting is problematic for a given adolescent, it could be indicative of other psychosocial problems.

Although some sexual behaviors can be risky (e.g., having multiple partners, unprotected sexual activity, concurrent substance use; Underwood et al., 2020), sexual exploration is a normative developmental process of adolescence (Harden, 2014; Kroger, 2006). When separated from sexual risk behaviors, sexual behavior in general is not consistently linked to poor psychological adjustment, especially without genetic or environmental confounds (Harden et al., 2008; Jamieson & Wade, 2011; Meier, 2007). Similarly, perhaps text-based sexting provides a normative outlet for adolescents to explore sexual intimacy (Campbell & Park, 2014; Holmes et al., 2020; Lippman & Campbell, 2014; Symons et al., 2018) and may be unrelated or positively related to adjustment. To understand comprehensively how sexting relates to adjustment, the present investigation examines parallel negative and positive adjustment outcomes, concurrently and longitudinally. We focus our literature review on studies that at least in part examined text-based sexting to better draw parallels between previous work and the current research.

Negative Adjustment Indicators

Text- and image-based sexting has almost exclusively been investigated with respect to negative adjustment, perhaps due to the widely held idea that adolescent sexting is inherently harmful (Best & Bogle, 2014; Englander, 2019). According to co-construction theory, if offline sexual behavior is linked to negative adjustment, sexting would share these same associations. Indeed, some evidence indicates that early sexual behavior may be a marker of poorer adjustment (Boislard-P. & Poulin, 2011; Lansford et al., 2014; McLeod & Knight, 2010; Scott-Sheldon et al., 2016; Whitbeck et al., 1999). However, these may only indicate concurrent adjustment-related issues rather than a causal relationship (Harden, 2014; Jamieson & Wade, 2011). The current research examined four common indices of negative adjustment: internalizing symptoms, externalizing symptoms, social problems, and borderline personality features.

Internalizing symptoms.

Text- and image-based sexting may be linked to internalizing difficulties, or broadly experiencing anxiety and/or depression-related issues. Although offline sexual behavior in adolescence is sometimes linked to heightened internalizing difficulties (McLeod & Knight, 2010; Whitbeck et al., 1999), this link is not consistently supported (Lehrer et al., 2006; Sabia, 2006) or could simply be a marker of previous or broader biopsychosocial issues (Jamieson & Wade, 2011; Meier, 2007). In line with co-construction theory, similar associations have been found between internalizing symptoms and sexting. Adolescents may experience or anticipate potential negative consequences of sexting (e.g., nonconsensual sharing) that generate anxiety or depression. One meta-analytic review found that sexting is positively related to internalizing symptoms (Mori et al., 2019). However, recent findings indicate nuance in this link—in a self-report study of consensual, non-consensual, and pressured sexting (defined as sending flirty or sexually explicit images or messages), consensual sexting was unrelated to depressive symptoms or self-harm behaviors (Wachs et al., 2021). Furthermore, few longitudinal investigations of sexting and internalizing symptoms exist; one study found that text- and image-based sexting was unrelated to depression/anxiety symptoms and self-esteem in a sample of Croatian adolescent girls followed over 20 months (Burić et al., 2020).

Externalizing symptoms.

Co-construction theory indicates that because there are positive relations between offline sexual behavior and externalizing issues (e.g., substance use; Boislard-P. & Poulin, 2011; Scott-Sheldon et al., 2016), sexting may similarly be related to externalizing symptoms. According to problem behavior theory, when adolescents violate rules in one way, they are likely to push boundaries in other ways. Text-based sexting, if problematic, may relate to externalizing symptoms through behavioral regulation problems, which also contribute to difficulties with impulsivity, aggression, or rule-breaking (Jiménez-Barbero et al., 2016). However, as offline sexual behavior may relate to externalizing issues due to confounding familial, genetic, or environmental factors (Donahue et al., 2013; Harden et al., 2008), text-based sexting may be unrelated to externalizing issues.

Social problems.

Text- and image-based sexting occurs within the context of social relationships and thus may affect adolescents’ social adjustment. Offline sexual behavior has been linked to greater social difficulties, such as rejection (Lansford et al., 2014), though this link has been inconsistently supported and subsequently challenged by more sex-positive research (Donahue et al., 2013; Harden, 2014). According to co-construction theory, sexting may also relate to greater social problems, though this may also be an artifact of negative perceptions of sexuality. Concurrent and longitudinal evidence also suggests relations between sexting and greater social difficulties (Gámez-Guadix & Mateos-Pérez, 2019; Ojeda et al., 2019; Van Ouytsel et al., 2019). For example, adolescents’ self-reported engagement in text- and image-based sexting was bidirectionally associated with cyberbullying engagement over one year (Gámez-Guadix & Mateos-Pérez, 2019) and both traditional bullying and cyberbullying across two years (Van Ouytsel et al., 2019). Perhaps, as posited by problem behavior theory, adolescents who push boundaries by engaging in text-based sexting may not be socially sophisticated and struggle in their social relationships. Or perhaps, as suggested by co-construction theory, poorer offline social relations are a result of negative social repercussions that may result from sexting (e.g., bullying or teasing about non-consensually shared sexts).

Borderline personality disorder features.

Sexual behaviors may also be positively related to borderline personality disorder (BPD) features, which are characterized by both emotional lability and volatile social relationships (Lieb et al., 2004). One longitudinal study of sexual risk behaviors and BPD symptoms in adolescent girls found positive links between the two, though only increased BPD symptoms at an earlier age predicted later sexual risk behaviors. In contrast, a previous analysis from this same longitudinal study found links between the frequency of text-based sexting in tenth grade and increased BPD features in emerging adulthood (Brinkley et al., 2017).

Positive Adjustment Indicators

To examine the way that text-based sexting relates to adolescents’ psychosocial adjustment beyond negative indicators, the current research also included indicators of positive adjustment that represent similar domains to the negative indicators. Sexual behavior is normative and necessary in adolescent development (Harden, 2014) and has been found to relate to greater well-being (Vrangalova & Savin-Williams, 2011) and social connectedness (Markham et al., 2010). In accord with co-construction theory, sexting could similarly be normative and be unrelated to or related to greater levels positive adjustment (Campbell & Park, 2014; Holmes et al., 2020; Lippman & Campbell, 2014; Symons et al., 2018). The current study examined life satisfaction, group belongingness, and three relevant dimensions of positive self-perceptions: social competence, behavioral conduct, and global self-worth.

Life satisfaction.

Few previous investigations have related text- and image-based sexting to overall measures of well-being such as life satisfaction. In accordance with both problem behavior and co-construction theory, if sexting is problematic, perhaps the distress or tension caused by engaging in this problematic behavior extends to adolescents being dissatisfied with life, or adolescents who are dissatisfied with life may seek the thrill of sexting. However, if text-based sexting is normative, it may make adolescents feel excited and mature, promoting their enjoyment of life concurrently and over time.

Group belongingness.

Group belongingness is characterized by strength of affiliation with an ingroup (e.g., Newman et al., 2007). Because text- and image-based sexting requires social interaction, it may be related to how adolescents perceive their own group belongingness. If text-based sexting violates social norms, engaging in this behavior alienates adolescents from their social connections, leading them to feel less like they belong. Alternatively, if text-based sexting is normative in their social groups, engaging in this behavior may allow adolescents to feel connected and valued, as suggested by similar relations seen with offline sexual behavior (Markham et al, 2010).

Self-concept.

According to co-construction theory, if adolescents can safely and effectively explore their sexuality via digital communication, they may generally view themselves positively. The dimensions of self-concept examined in the current investigation reflect relevant aspects of psychosocial adjustment, namely social competence, behavioral conduct, and self-worth. As noted above, text- and image-based sexting is linked with social difficulties (e.g., traditional and cyber-victimization; Gámez-Guadix, & Mateos-Pérez, 2019; Ojeda et al., 2019; Van Ouytsel et al., 2019), suggesting that those who sext may report lower levels of self-perceived social competence. However, given links between offline sexual activity and social connectedness (Markham et al., 2010), adolescents who engage in text-based sexting may report stronger social competence.

Behavioral conduct is characterized by the degree to which an adolescent perceives themselves as engaging in appropriate behavior and avoiding trouble (Harter, 1988). The previously noted links between text- and image-based sexting and externalizing problems suggest that those who sext may report poorer self-perceived behavioral conduct. Perhaps adolescents who sext may be aware that they are less likely than peers to act appropriately. However, if text-based sexting is normative, this may not be the case.

Finally, text-based sexting may be linked to greater global self-worth because it can be exciting and make adolescents feel more mature. Previous research examining image-only sexting and image- and text-based combined sexting found no relation to self-esteem (Gordon-Messer et al., 2013; Klettke et al., 2019); however, given that text-based sexting may be qualitatively different than image-based sexting, its relation with global self-worth must be examined separately.

Gender Differences in Adjustment Associations

Although previous research has not found strong gender differences in the prevalence of sexting (Madigan et al., 2018), the associations between text-based sexting and adjustment may differ by gender. The framework of gendered double standards indicates that men/boys and women/girls are held to different expectations regarding engagement in sexual behavior and sexting specifically (Dir et al., 2013; Lippman & Campbell, 2014; Peréz Domínguez, 2020; Ringrose et al., 2013; Walker et al., 2013). Girls may both be expected to demonstrate their sexuality and negatively affected by sexting behaviors given that they are often scrutinized for sexual expression; conversely, boys are permitted or even encouraged and praised for engaging in sexual behavior, including sexting (Dir et al., 2013; Thorburn et al., 2021).

Further, the gendered nature of sexting can be considered within the frame of objectification theory, which posits that women are socially valued as physical or sexual objects (Fredrickson & Roberts, 1997; Moradi & Huang, 2008). This process is evident even for girls as young as 13, though the effects seem to strengthen with age (for a review, see Daniels et al., 2020). Objectification and sexualization have been understood to have negative consequences for women’s and girls’ self-concept, whereby they learn to prioritize sexual attractiveness and critique their own image and behaviors (Moradi & Huang, 2008). Applying objectification theory to sexting suggests that engagement in sexting may uniquely affect girls. Girls may be socialized to understand themselves as sexual beings and therefore be compelled to engage in sexual behaviors such as sexting. Indeed, self-objectification was found to relate to positive sexting attitudes and behaviors in adolescent girls (Bianchi et al. 2017; Speno & Aubrey, 2019). The pressure to perform as sexual beings may confer psychosocial problems such as depressive symptoms for girls (Grabe et al., 2007); therefore, sexting may indicate particularly negative psychosocial adjustment outcomes for girls relative to boys. However, in contrast to objectification theory, sexting may allow for girls to explore their sexuality in a physically safer way than offline sexual behavior, which may not be indicative of psychosocial adjustment (Burić et al., 2020; Ringrose et al., 2013).

As sexting operates within a gendered and sexualized social context, gender differences may exist in the association between text-based sexting and psychosocial adjustment. The potential dangers of solely image-based sexting for girls have been investigated (cf. Ringrose et al., 2013; Woodward et al., 2017); few studies of both text-based and image-based sexting examine gender moderation. In one study, text- and image-based sexting was related to higher cyberbullying victimization in girls compared to boys (Barrense-Dias et al., 2017), but in a sample of girls, text- and image-based sexting was unrelated to internalizing symptoms (Burić et al., 2020). In a meta-analytic review, gender did not emerge as a moderator of the link between sexting and mental health or risky behaviors (Mori et al., 2019).

The Present Study

This study investigated adolescents’ observed text-based sexting in relation to both negative and positive aspects of psychosocial adjustment. To explore sexting’s role in developmental processes and outcomes, it is necessary to examine relations for adolescents who engage in text-based sexting more than others on average (between-person differences) and at times when adolescents engage in text-based sexting more than they do on average (within-person differences) over time. This study examined the sending and receiving of text-based sexts and explored four questions: (1) Do adolescents who text-based sext more than others overall exhibit more positive or more negative adjustment (i.e., concurrent between-person differences)? (2) When adolescents text-based sext more at one time-point than they do on average, do they exhibit more positive or more negative adjustment (i.e., concurrent within-person differences)? (3) Does text-based sexting more than one’s own average predict later adjustment or vice versa (i.e., within-person longitudinal associations)? and (4) Are there gender differences in the associations between text-based sexting and adjustment outcomes, concurrently or longitudinally?

For the first three questions, we posed two general competing hypotheses. If text-based sexting is indicative of problematic behavior, according to problem behavior theory, it should be related to more negative adjustment, including higher levels of negative adjustment indicators and lower levels of positive adjustment indicators. However, if text-based sexting aids in normative development and allows adolescents to cultivate their psychosocial and sexual health, it should be related to more positive adjustment, including lower levels of negative adjustment indicators and higher levels of positive adjustment indicators. The fourth question was exploratory and intended to investigate potential effects of the gendered and objectifying context of sexting behavior.

Methods

Participants

Participants were 197 typically developing adolescents (48% female, M_age= 13.93 during 9^th grade) from several high schools in the Southwestern United States, taking part in a longitudinal study of social aggression and digital communication. Ninety-five percent of the sample identified as heterosexual in the 12^th grade (i.e., chose “Straight” when prompted with “Do you consider yourself to be…” and provided with the options “Straight,” “Gay/lesbian,” “Bisexual,” or “Questioning”). The sample was 53% White, 22% Black, 18% Hispanic, and 7% other/mixed race. Inclusion in these analyses required data for at least two time-points of text message content and self-report data; eighteen participants were therefore excluded from the original sample (N = 215) to result in a total of 197 participants for this study.

Procedures

When participants were under age 18, they provided annual assent and parents provided annual consent to participate in the study; participants provided consent after turning 18. All study procedures were approved by The University of Texas at Dallas Institutional Review Board (IRB#: 07–36). Data for the current study were collected from Fall 2009 through Summer 2012. Participants completed self-report surveys about their psychosocial adjustment yearly in the summers after 9^th through 12^th grades (or approximately 14 to 18 years old).

Participants were given smartphones with paid service plans including unlimited texting and data usage prior to 9^th grade. Participants were encouraged, but not required, to use the study-provided smartphone as their primary communication device. Most participants reported that they “almost always” utilized their study-provided smartphone, although they were allowed to use other devices and some reported “sometimes” texting on another phone (Underwood et al., 2012). Despite being monitored and able to use other phones, examination of the captured data suggests similar rates of sexual and profane language to those found in unmonitored online chat rooms (Subrahmanyam et al., 2006). Participants could delete messages from their phones if desired, so they had control of the privacy of their physical smartphone device. The research team monitored participants’ text messages for indications of harm to self or others; instances were reported to the necessary authorities. Participants were informed annually at the time of assent that messages would be monitored; 90% reported that they did not adjust their behavior as a result (Meter et al., 2019). A description of the full methodology for text message capturing and ethical considerations can be found elsewhere (Underwood et al., 2012; Underwood et al., 2015).

The smartphones archived the text content of all incoming and outgoing messages across participants’ four years of high school. Due to the wealth of archived text messages (over 27 million text messages across four years), it was not practical to code all text messages, so a sample of days was selected. Two, two-day-long transcripts of all incoming and outgoing text messages were selected for each participant (two consecutive days in the fall and two consecutive days in the spring of each academic year); thus, participants had up to a total of sixteen days of digital communication. For each year, the days in the fall were selected around Homecoming events and the days in the spring were selected around Valentine’s Day, as it was expected that communication with peers might have increased during these periods. Transcripts were coded by a team of undergraduate research assistants who were trained over the course of a semester. Each text message was coded by utterance, or phrase that conveys a complete thought (e.g., “we should hang out”, “I’m in class”). Utterances also were determined to have agents and targets, defined as who is initiating/directing the utterance (agent) and who is receiving it or for whom it is intended (target). A reliability coder coded 20% of all text message transcripts, and coders met weekly to discuss challenging codes and maintain coding proficiency.

Measures

Sexting

Sexual talk was coded as content about sexual behaviors past or presently occurring and hypothetical sexual behaviors (κ’s=.69-.80, in the substantial range according to Landis & Koch, 1977). Sexual talk included plans or desires to engage in sexual behaviors (e.g., “You should come over after school so we can have sex”), as well as recounting previous sexual behaviors (e.g., “Remember when we did it after the party?”). Current arousal (e.g., “You’re making me hot”) or masturbation (e.g., “My fingers are wandering…”) were also coded as sexual talk. Our archiving software was unable to capture pictures; however, asking for or indications of sending sexually explicit pictures (e.g., “Thanks for that sexy pic”) were coded as sexual talk.

Text-based sexting was defined as sexual talk that occurred “in-dyad,” or where both the agent and target of the utterance were one of the two people communicating in the text message conversation. Therefore, utterances were coded as sexting if they were directed toward the dyad partner (e.g., “I want to hook up with you”) or reflected solo sexual behavior that one of the dyad partners was engaging in (e.g., “I’m touching myself”). The coding system captured all sexual talk that involved any person outside the dyad, but these types of utterances were not included in the current definition of text-based sexting (e.g., “I’m planning to go all the way with Alex;” “Taylor and Jordan fooled around last summer”). The focus on in-dyad sex talk was made to ensure that the observation of text-based sexting was more comparable to previous conceptualizations of sexting in self-report studies (cf. Mori et al., 2019), as opposed to flirting, bragging about sex with others, or gossiping about others’ sexual behavior. The proportion of sex-related talk captured by in-dyad sexting is reported in Supplemental Table S1, and analyses with all sex talk are included in the Supplemental Tables S4 and S5, to provide comparisons between this overall sex talk behavior with the focused data that are used in the current investigation.

The number of text-based sexts observed in each transcript was summed to yield a total number of text-based sexts per transcript, and was then summed across the four transcripts within a given year to yield a total number of text-based sexts per four-day period for each academic year (ninth, tenth, eleventh, and twelfth grade). This summary was created to approximate an annual assessment, to parallel the annual assessment of adjustment variables during the summer after each academic year. Although self-reports and texting observations were separated by several months, we conservatively consider them within-year and concurrent. Sexts were initially separated into sexts sent by the participant and sexts received by the participant, though these were combined given an extremely high correlation.

Gender

As part of a basic demographic information questionnaire, participants were queried “What is your gender?” and provided with binary forced-choice options of male or female. Participants were likely to interpret this question as gender and not sex assigned at birth, given the explicit phrasing.

Negative Adjustment

Participants completed the 112-item Achenbach System of Empirically Based Assessment Youth Self-Report (Achenbach & Rescorla, 2001), which broadly measures youth psychosocial adjustment “now or within the past 6 months” on a 0–2 scale (0=Not true, 2=Very true or often true). The internalizing scale sums the anxious-depressed, withdrawn-depressed, and somatic complaint syndrome scales (31 items, e.g., “I feel overtired without good reason; α’s=.87-.91) The externalizing scale sums the aggressive behavior and rule-breaking syndrome scales (32 items, e.g., “I am mean to others”; α’s=.85-.91). The social problems syndrome scale includes 11 items (e.g., “I don’t get along with other kids”; α’s=.72–83). Participants also completed the McLean Screening Instrument for Borderline Personality Disorder (MSI-BPD; Zanarini et al., 2003), in reference to the past year, with items in yes/no format (10 items, e.g., “Have you been extremely moody?”; α’s=.74-.80).

Positive Adjustment

Participants’ total life satisfaction was assessed by summing items measuring their happiness at school, outside of school, and their whole life (e.g., “I am happy about my life at school”), using a 1–5 scale (1=Never, 5=All of the time; α’s=.79-.84); higher scores indicated higher life satisfaction.

Participants’ feelings of belongingness within their friend group were assessed using the 14-item Group Belonging scale (Newman & Newman, 1993) using a 1–5 scale (1=Not at all true, 5=Very true; e.g., “I am proud to be part of my friendship group”; α’s=.83-.85); higher scores indicate higher belongingness.

In the summers after 9^th through 11^th grade only, participants completed the 45-item Self-Perception Profile for Adolescents (Harter, 1988) which measures various facets of self-concept. Participants were presented with items that contain two opposing statements, one positive and one negative. Participants chose which statement best represented how they felt about themselves and selected whether it was “Really true of me” or “Sort of true of me”. The current research examined scales (five items each) assessing social competence (e.g., “Some teenagers find it hard to make friends BUT For other teenagers it’s pretty easy”; α’s=.82-.83), behavioral conduct (e.g., “Some teenagers usually do the right thing BUT Other teenagers often don’t do the right thing”; α’s=.77-.83), and global self-worth (e.g., “Some teenagers are often disappointed with themselves BUT Other teenagers are pretty pleased with themselves”; α’s=.83-.85). Higher scores for each scale indicate higher levels of that self-perception.

Analysis Plan

Descriptive analyses were conducted in IBM SPSS, Version 24.0. Differences in sexting by gender and by grade and the interaction between gender and grade were determined with two-way ANOVAs. To test the within-year (concurrent) and longitudinal associations between sexting and adjustment, a series of multilevel models were conducted in Mplus8 (Muthén & Muthén, 2017). Each model was conducted separately using the full sample. Multi-group analyses were conducted to test for gender moderation and cross-level interaction coefficients were examined to compare the relative strength in associations (i.e., beta coefficients) between boys and girls. There was a substantial amount of missing data in sexting (12% missing) and adjustment (ranged from 35%, because the self-perception scales were not collected at the final time point, to 6% missing). To account for non-normality, missing data, and keep the sample size consistent across model type (690 data points for concurrent models and 788 data points for longitudinal models), models were estimated using full information maximum likelihood estimation with robust standard errors. Participants’ average total texting was added as a covariate to all models. Given the numerous comparisons, significance values were adjusted using the Benjamini-Hochberg correction for each research question to adjust for false discovery rates (Benjamini & Hochberg, 1995).

$$\begin{array}{ll}\text{Level}1:\hfill & {\text{Adjustment}}_{\text{ij}}={\mathrm{\beta }}_{\text{0j}}+{\mathrm{\beta }}_{\text{1}}\left({\text{wSexting}}_{\text{ij}}\right)+{\mathrm{\beta }}_{\text{2}}\left({\text{Grade}}_{\text{ij}}\right)+{\mathrm{\epsilon }}_{\text{ij}}\hfill \\ \text{Level}2:\hfill & {\mathrm{\beta }}_{\text{0j}}={\mathrm{\gamma }}_{\text{00}}+{\mathrm{\gamma }}_{\text{01}}\left({\text{wSexting}}_{\text{j}}\right)+{\mathrm{\gamma }}_{\text{02}}\left({\text{mTotText}}_{\text{j}}\right)\hfill \\ \hfill & +{\mathrm{\gamma }}_{\text{03}}\left({\text{Male}}_{\text{j}}\right)+{\mathrm{\upsilon }}_{\text{0j}}\hfill \end{array}$$ Equation 1

Equation 1 shows the within-year multilevel models testing the within- and between-person sexting variables’ associations with adjustment. Within-person sexting variables were created by subtracting the individual participant’s sexting mean across all four years from the amount of sexting at each grade level (Curran & Bauer, 2011). These yearly deviations in sexting (Level 1, β₁; wSexting_ij) allowed us to examine the within-person associations with adjustment. Between-person variables were calculated by subtracting the grand mean of sexts from each participant’s average sexting rates (Curran & Bauer, 2011). This mean sexting difference (Level 2, γ₀₁; mSexting_ij) allowed us to examine between-person associations with adjustment. Models also accounted for initial individual differences in sexting with a random intercept (β_0j), participants’ average total texting frequency with γ₀₂(mTotText_j), participants gender (γ₀₃(Male_j)), residual term (ε_ij), and a random person-specific error term (υ_0j). Grade (β₂; Grade_ij), the index for time, was treated linearly (resulting in the best model fit as compared to categorical and quadratic models) and centered such that the intercept indicated levels at 9^th grade.

$$\begin{array}{ll}\text{Level}1:\hfill & {\text{Adjustment}}_{\text{ij}}\text{or}{\mathrm{\beta }}_{\text{0j}}+{\mathrm{\beta }}_1\left({\text{wSexting}}_{\text{i}-1}\right)+{\mathrm{\beta }}_{\text{2}}\left({\text{wAdjustment}}_{\text{i}-1}\right)\hfill \\ \hfill & {\text{Sexting}}_{\text{ij}}=+{\mathrm{\beta }}_3\left({\text{Grade}}_{\text{ij}}\right)+{\mathrm{\epsilon }}_{\text{ij}}\hfill \\ \text{Level}2:\hfill & {\mathrm{\beta }}_{\text{0}}={\mathrm{\gamma }}_{\text{00}}+{\mathrm{\gamma }}_{\text{01}}\left({\text{wSexting}}_{\text{j}}\text{or}{\text{mAdjustment}}_{\text{j}}\right)+{\mathrm{\gamma }}_{\text{02}}\left({\text{mTotText}}_{\text{j}}\right)\hfill \\ \hfill & +{\mathrm{\gamma }}_{\text{03}}\left({\text{Male}}_{\text{j}}\right)+{\text{υ}}_{\text{0j}}\hfill \end{array}$$ Equation 2

Equation 2 provides the equations for the autoregressive, cross-lagged multilevel models used to test for bidirectional, longitudinal associations between sexting and adjustment. These models also included covariates (grade, gender, average total texting), nesting within individuals (random intercept), and error terms. The models also accounted for the between-person means of sexting (γ₀₁; mSexting_j) and adjustment (γ₀₁; mAdjustment_j) in level 2. The models included longitudinal cross-lagged within-individual paths simultaneously testing associations from sexting to later adjustment (β₁; wSexting_i-1) and adjustment to later sexting (β₁; wAdjustment_i-1), after accounting for the autoregressive paths—the previous rates of adjustment (β₂; wAdjustment_i-1) and sexting (β₂; wSexting_i-1).

Results

Table 1 presents descriptive statistics for total texts, total text-based sexts sent, and total text-based sexts received, for the full sample, by grade, and by gender; correlations among sexting and the adjustment indicators, averaged across the four years, are reported in Table S2 in the Supplemental Information. On average, participants sent and received about 5 sexts during any one of the four-day sample periods throughout high school. Sent and received sexts were highly correlated (r=.93), and 40 percent of participants sent or received at least one sext at some point during the 16-day sampling period. The frequency of sending and receiving sexts did not differ by grade. No gender differences emerged in amount of sexting engagement.

Table 1.

Descriptive information of texting and sexting variables for the full sample, by grade, and by gender.

Statistic	Full Sample	Grade						Gender
		9th	10th	11th	12th	F	p	Boys	Girls	F	p
Total Texts
M	611.42	424.03a	746.86b	669.18bc	585.23c	10.42	<.001	618.00a	604.10a	.11	.75
SD	559.81	388.63	664.07	649.68	416.33			645.56	446.19
Range	1–5619	3–2106	2–5619	21–4453	1–2380			1–5619	2–2810
Skew	2.85	1.81	2.94	2.75	1.41			3.16	1.25
In-Dyad Sexts
M	11.11	7.59a	14.96a	13.27a	9.22a	2.18	.10	13.44a	8.51b	3.97	.05
SD	32.51	18.61	43.28	38.70	20.10			38.10	24.59
Range	0–389	0–106	0–389	0–356	0–162			0–389	0–273
Skew	6.67	3.54	5.79	5.99	4.80			6.14	6.77
% who sent sexts	41.22	29.94	43.26	44.07	46.33			40.11	42.46
% who received sexts	42.59	35.03	45.51	41.80	47.16			46.01	38.77

Note. Shared superscript indicates no statistically significant mean difference (via pairwise comparisons using Bonferroni corrections; analyzed separately by grade and by gender). % who sexted indicate those who sent and/or received at least one sext during any of the selected two-day periods.

Between-Person Differences in Sexting and Adjustment

Results of the concurrent multilevel models are presented in Table 2. Engaging in sexting more on average than one’s peers was not associated with any measure of adjustment after false discovery rate correction (p’s>.03, ns after Benjamini-Hochberg correction).

Table 2.

Unstandardized and standardized coefficients, standard errors, and p-values of the concurrent within- and between-person associations between sexting and adjustment indicators.

Adjustment Indicators	Sexts
	Between-Person Differences				Within-Person Differences
	b	SE	β	p	b	SE	β	p
Internalizing Problems	0.06	0.03	0.22	0.03	−0.02	0.01	−0.07	0.07
Externalizing Problems	0.08	0.05	0.26	0.10	0.002	0.01	0.01	0.91
Social Problems	0.01	0.01	0.14	0.08	−0.01	0.004	−0.06	0.20
Borderline Personality Symptoms	0.02	0.01	0.26	0.13	−0.004	0.003	−0.06	0.19
Life Satisfaction	−0.004	0.004	−0.16	0.31	−0.001	0.001	−0.03	0.52
Group Belongingness	−0.003	0.003	−0.17	0.28	−0.001	0.001	−0.08	0.003
Social Competence	−0.003	0.003	−0.13	0.33	0.001	0.001	−0.01	0.88
Behavioral Conduct	−0.003	0.003	−0.12	0.31	−0.001	0.001	−0.07	0.13
Global Self-Worth	−0.001	0.003	−0.06	0.63	0.001	0.001	0.03	0.47

Note: Separate models were conducted for each adjustment variable, with between- and within- variables included simultaneously in each model. Models controlled for gender, grade, and participants’ average total texting. Bolded font indicates significance after FDR correction.

Within-Year, Within-Person Differences in Sexting and Adjustment

Table 2 also presents the within-year, within-person associations between sexting and adjustment. After correction, engaging in sexting more at one assessment point than one’s average was negatively related to group belongingness. No other relations emerged for the other adjustment outcomes (p’s>.07).

Longitudinal Associations between Sexting and Adjustment

Results of the bidirectional longitudinal multilevel models are presented in Table 3. After accounting for family-wise error rate correction, text-based sexting more than one’s own average in one year did not predict adjustment in the following year, nor did adjustment predict the amount of text-based sexting in the following year (p’s>.006, ns after Benjamini-Hochberg correction).

Table 3.

Unstandardized and standardized coefficients, standard errors, and p-values of the longitudinal associations between sexting and adjustment indicators.

Adjustment Indicators	Sexts
	Lagged Sext Adj				Lagged Adj Sext
	b	SE	β	p	b	SE	β	p
Internalizing Problems	0.02	0.02	0.10	0.13	−0.38	0.34	−0.07	0.26
Externalizing Problems	0.02	0.01	0.11	0.09	−0.34	0.40	−0.06	0.39
Social Problems	0.01	0.01	0.14	0.01	−1.15	0.76	−0.08	0.13
Borderline Personality Symptoms	0.001	0.003	−0.004	0.93	−0.79	1.25	−0.04	0.53
Life Satisfaction	0.001	0.001	−0.02	0.79	6.15	3.58	0.09	0.09
Group Belongingness	−0.001	0.001	−0.08	0.32	7.81	4.52	0.10	0.08
Social Competence	−0.004	0.001	−0.24	0.006	12.18	6.47	0.15	0.06
Behavioral Conduct	0.00	0.001	0.03	0.52	−4.74	3.32	−0.06	0.15
Global Self-Worth	0.00	0.001	−0.02	0.72	2.75	3.70	0.03	0.46

Note: Separate models were conducted for each adjustment variable, but sexting and adjustment longitudinal paths were modeled simultaneously. Models controlled for between-person means of sexting and adjustment, autoregressive paths, gender, grade, and average total texting. No associations were significant after FDR correction.

Gender Differences in Sexting and Adjustment Associations

Table 4 contains the between-person and within-year, within-person associations separated by boys and girls. For girls, sexting more than others was related to increased internalizing problems, externalizing problems, and borderline personality disorder features, and decreased life satisfaction, group belongingness, self-perceived social competence, and global self-worth. Cross-level interaction coefficients revealed differences in the associations between sexting and adjustment such that girls demonstrated a stronger between-person relations for sexting and externalizing problems (b=−.204, p=.002), borderline personality disorder features (b=−.051, p<.001), life satisfaction (b=.016, p<.001), group belongingness (b=.009, p=.001) self-perceived social competence (b=.009, p=.01), behavioral conduct (b=.012, p<.001) and global self-worth (b=.011, p<.001). For boys, sexting more at one time point than one did on average was related to decreased social problems; however, boys did not demonstrate a stronger relation than girls (b=−.03, p=.15). After correction, no other concurrent between- or within-person associations emerged for boys or girls. Similarly, after correction, longitudinal models revealed no significant gender differences in the lagged associations (Table S3; ps>.02, ns after Benjamini-Hochberg correction).

Table 4.

Coefficients, standard errors, and p-values of the concurrent associations between sexting and adjustment indicators by gender.

Adjustment Indicators	Sexts (Girls)
	Between-Person Differences				Within-Person Differences
	b	SE	β	p	b	SE	β	p
Internalizing Problems	0.11	0.04	0.28	0.005	−0.001	0.03	−0.002	0.10
Externalizing Problems	0.19	0.07	0.42	0.009	−0.003	0.02	−0.009	0.91
Social Problems	0.03	0.01	0.16	0.04	0.005	0.01	0.04	0.65
Borderline Personality Symptoms	0.05	0.009	0.51	<.001	−0.001	0.004	−0.02	0.74
Life Satisfaction	−0.02	0.002	−0.43	<.001	−0.002	0.002	−0.07	0.30
Group Belongingness	−0.01	0.002	−0.36	<.001	−0.002	0.001	−0.08	0.006
Social Competence	−0.01	0.003	−0.32	<.001	0.002	0.002	0.08	0.37
Behavioral Conduct	−0.008	0.004	−0.22	0.06	−0.001	0.001	−0.03	0.65
Global Self-Worth	−0.008	0.003	−0.26	0.004	0.002	0.001	0.10	0.11
Adjustment Indicators	Sexts (Boys)
	Between-Person Differences				Within-Person Differences
	b	SE	Β	p	b	SE	β	p
Internalizing Problems	0.04	0.03	0.19	0.25	−0.03	0.009	−0.11	0.009
Externalizing Problems	0.03	0.04	0.11	0.49	0.005	0.02	0.03	0.73
Social Problems	0.01	0.01	0.20	0.18	−0.01	0.003	−0.12	0.001
Borderline Personality Symptoms	−0.001	0.008	−0.004	0.98	−0.005	0.004	−0.08	0.19
Life Satisfaction	0.002	0.003	0.07	0.61	−0.001	0.001	−0.004	0.94
Group Belongingness	0.001	0.003	0.04	0.83	−0.001	0.001	−0.09	0.03
Social Competence	0.001	0.003	0.01	0.93	−0.001	0.001	−0.08	0.10
Behavioral Conduct	−0.001	0.002	−0.08	0.58	−0.001	0.001	−0.07	0.21
Global Self-Worth	0.002	0.002	0.10	0.44	0.001	0.001	0.005	0.92

Note: Separate models were conducted for each adjustment variable, with between- and within- variables included simultaneously in each model. Models controlled for grade and average total texting. Bolded font indicates significance after FDR correction.

Discussion

This observational study examined concurrent and longitudinal associations between adolescents’ text-based sexting and indicators of both negative and positive adjustment. When considering boys and girls together, text-based sexting was not significantly related to adjustment outcomes, with the exception of group belongingness. However, notable findings emerged when examining associations by gender: girls, but not boys, who engaged in more text-based sexting than their peers also reported significantly poorer adjustment for several outcomes. No longitudinal associations emerged. Thus, although text-based sexting may not be emblematic of problems for adolescents overall, girls who engage in more text-based sexting than their peers may be experiencing greater psychosocial difficulties.

Sexting and Psychosocial Adjustment

Within-Year Associations between Sexting and Adjustment

Between-Person.

For the full sample, text-based sexting more than one’s peers was largely unrelated to adjustment. When considering boys and girls separately, the between-person relations between text-based sexting and adjustment were stronger for girls. Girls’ text-based sexting was associated with higher externalizing problems and borderline personality disorder features, and lower life satisfaction, group belongingness, self-perceived social competence, and global self-worth.

This pattern of results suggests that for girls, engaging in text-based sexting in excess relative to one’s peers may be concomitantly associated with psychosocial difficulties. The few previous investigations of sexting as it relates to mental and social health for girls have found no evidence of unique associations between sexting and psychosocial health for girls (e.g., Barrense-Dias et al., 2017; Burić et al., 2020; Mori et al., 2019). However, it is increasingly recognized that boys and girls are held to different expectations for engaging in sexting (Dir et al., 2013; Lippman & Campbell, 2014; Peréz Domínguez, 2020; Walker et al., 2013). Perhaps girls experience this as a double-standard pressure to both sext and not sext (Lippman & Campbell, 2014; Ringrose & Gill, 2013), which may be associated with negative adjustment when girls do engage in text-based sexting, especially in excess. Additionally, perhaps objectification, where girls’ social value is placed upon their sexual attractiveness, contributes to some girls’ engagement in text-based sexting at higher rates than their peers and is simultaneously associated with adjustment problems. Indeed, objectification has been well understood to lead to negative adjustment such as internalizing symptoms (Fredrickson & Roberts, 1997; Tiggemann & Slater, 2015). Further, maybe excess text-based sexting is a marker of stronger pressures for girls to sext, which, according to co-construction theory, may reflect broader social difficulties such as a lack of social connectedness and perceived social skills. In accordance with problem behavior theory, perhaps higher-than-normative text-based sexting for girls indicates other behavioral or emotional regulation issues. This is reflected in the relations of text-based sexting more than one’s peers with externalizing problems and borderline personality disorder features.

Future research should investigate the mechanisms underlying the gender differences in the associations between text-based sexting and adjustment. Further examinations of objectification in the context of sexting behaviors are sorely needed. Additionally, perhaps the content and context of sexting conversations are especially important for understanding how sexting may relate to adjustment differently for boys and girls. For example, are gendered double-standards and objectification of girls actually emerging in sexting conversations? In other words, are girls receiving pressures to sext and shame when they do sext? Further observational methods and content analyses could elucidate these patterns and trends in sexting conversations to understand how they may relate to adolescents’ psychosocial health.

Within-Person.

When considering text-based sexting more at one time point than one does on average, only one association with adjustment emerged: higher text-based sexting at one time point than one’s own average was negatively related to group belongingness. No significant differences in the within-person associations emerged between boys and girls. These results suggest that, regardless of gender, higher-than-usual amounts of text-based sexting may be acutely related to a lack of social connectedness. According to co-construction theory, adolescents may be trying to construct intimate or provocative online social interactions to counter offline social loneliness. Perhaps when adolescents feel disconnected to a group of peers, they may exchange text-based sexts with a partner to feel less isolated. Alternatively, perhaps when adolescents send more text-based sexts than they normally do, their peers notice and react to this difference. Peers could be withdrawing from adolescents who are engaging in higher-than-usual amounts of text-based sexting in response to this potential violation of social norms, or perhaps adolescents who engage in higher-than-usual amounts of text-based sexting may be in new romantic involvements and consequently spending less time with their friends. Overall, the lack of associations when considering text-based sexting relative to one’s own average suggests that engaging in sexting in general may be normative—sexting more than others (not more than one’s own average) may be what reflects problems, specifically for girls.

Longitudinal Associations between Sexting and Adjustment

No longitudinal associations with text-based sexting emerged in any analysis. Though sexting has been conceptualized as developmentally harmful (Best & Bogle, 2014; Englander, 2019), it seems that text-based sexting may be part of adolescents’ normative sexual and psychosocial development. Scholars are beginning to conceptualize sexting as a nuanced and sometimes normative process (Campbell & Park, 2014; Holmes et al., 2020; Lippman & Campbell, 2014; Symons et al., 2018); perhaps it is more important to attend to the frequency of sexting behavior relative to an adolescent’s peers.

Limitations

This study had several strengths, including the longitudinal design, observational methods, inclusion of within- and between-person differences, and examination of gender differences. However, findings also need to be considered in the context of several limitations. First, this study examined only amount of sexting. Future research should examine the specific context and content of sexting, such as whether it occurs within a romantic relationship, if it is reciprocal, or if there are elements of coercion. Second, the data were collected from 2008 to 2012; digital communication technologies and platforms have evolved since then (Pew Research Center, 2018). However, texting remains and has grown as a vital and preferred method of communication among contemporary adolescents, even compared to social media apps (Bailey et al., 2016; Common Sense Media, 2018).

Third, given the impossibility of content coding each of the over 27 million text messages archived in this study, the research team had to select sample days somewhat arbitrarily. Though the days were a-priori selected to maximize the likelihood of peer communication, these sample days may not wholly reflect the texting (or sexting) habits of these adolescents across the four years of the study. Relatedly, we averaged across sample days for each year to approximate yearly assessments of texting, to parallel the yearly assessments of adjustment. Given that these were not collected at the exact same time point, concurrent associations were actually within-year associations. To not overstate potential longitudinal relations, we conservatively considered these within-year associations to be concurrent. Additionally, as the larger study was broadly focused on social communication, we could not feasibly conduct more fine-grained coding of the content and context of text-based sexts, and re-coding the 468,201 observations for context within sexual conversations was not possible. Fourth, participants were allowed to use other devices for the duration of the study. Although participants reported using the study-provided phone “most of the time” (Underwood et al., 2012) and rates of profane and sexual talk mirrored those found in unmonitored online chat rooms (Subrahmanyam et al., 2006), it is possible that we did not capture all the sexual talk that occurred on each sample day.

Fifth, as this is the first study to observationally measure sexting as text-based discussions about sex between conversation partners, this study was limited to text-based sexting. Future research is needed that directly compares text-based sexting to image-based sexting and, given that their associated risks may differ, their relative associations with psychosocial adjustment. Relatedly, to understand whether the associations are specific to text-based sexting, future research should directly compare text-based sexting to in-person sexual talk or sexual talk exchanged over social media.

Finally, we only considered binary gender identity (boys/girls, assessed as “male” or “female”) in an overwhelmingly heterosexual sample. These data were collected from 2008 to 2012, when sex and gender terms were used interchangeably in developmental psychology research and before it was customary to assess gender using open-ended or multiple-choice response options, or providing options for non-binary, genderqueer, or transgender identities. Because this was the latter part of a longitudinal study starting in 2003 which required annual in-person contact, participants established a personal bond with the study personnel and it is likely that study personnel would have been informed by participants if they identified as transgender or nonbinary. Nevertheless, given that sexting occurs in a gendered context, gender and sexual minority identity experiences could affect the relations between sexting and psychosocial adjustment for adolescents. Future research should investigate not only the amount but the content and context of sexting among LGBTQ+ adolescents, using expansive measurement options.

Conclusions and Clinical Implications

The findings from the current study have important implications for research, policy, and clinical practice. First, the finding that higher within-person text-based sexting relates to lower belongingness may suggest that adolescents may sext with a partner when they feel disconnected or excluded from groups of peers. Alternatively, when adolescents may begin sexting more as part of a new romantic relationship, they may also spend less time with peers and thus feel disconnected from the group. Future research should investigate this link between lack of social belonging and text-based sexting. Practitioners and policy makers might focus on supporting adolescents in coping with feeling like one does not belong, such as reaching out to a trusted friend to discuss those feelings, finding affinity spaces online for adolescents’ other interests, or using interpersonal skills to connect with peers in a more fulfilling way.

Second, the unique between-person differences in text-based sexting for girls should be given more attention. Research should continue to explore the ways that sexting could be normative or non-normative for adolescents overall, as well as in the context of gender, to identify any problems with sexting and the processes by which they emerge. Practitioners and policy makers should consider the ways that sexting operates in gendered contexts and therefore may demonstrate divergent relations by gender with adolescents’ psychosocial health. For example, sexual health education efforts could situate discussions of normative and non-normative sexting behaviors in the context of gendered double standards and objectification, to help adolescents and caregivers identify when gendered double standards and objectification are influencing (and perhaps negatively impacting) their sexting behavior and wellbeing. Clinically, intervention attention toward ameliorating the difficulties associated with text-based sexting should be tailored to adolescents’ experiences of gender and gender-related pressures around sexting. For instance, clinicians can assess and explore adolescents’ gender identity, roles, and expectations around sexual behaviors, to identify ways that adolescents may safely engage in sexual exploration via sexting.

Contrary to many parents’ and clinicians’ fears that sexting is harmful to adolescent development, the present study overall found few concurrent or longitudinal associations between text-based sexting and adjustment problems in this typically-developing sample of adolescents. Future research and clinical efforts instead should be made to understand the nuances of sexting (e.g., types of sexting content) and whether sexting may reflect or perpetuate gendered experiences around girls’ sexual behavior.