With the advent and introduction of new communication technologies come debates about their utility and morality (Katz, Rice, Aspden, 2001). As Green and Clark (2015) explained, “Such debates arose first in the wake of the telegraph and have arisen anew from the development of every new technology since” (p. 247). One recent debate surrounds the use of digital devices, such as computers and mobile phones, to create and exchange messages and images of a sexual nature, a practice commonly referred to as sexting (Ringrose, Gill, Livingstone, Harvey, 2012). Two discourses dominate this debate, both in the academic literature and in the popular press (Doring, 2014; Rice et al., 2014). The prevailing discourse frames sexting as inherently risky, a deviant behavior in need of intervention and prevention. The competing frame positions sexting as a normal, even healthy aspect of sexual expression and relationships (Doring, 2014). Evidence in support of this normalcy perspective is growing (Cooper, Quayle, Jonsson, Svedin, 2016), but far more sexting studies focus on its potential negative consequences and links to problematic behaviors. In fact, Doring (2014), in a review of 50 academic papers on sexting, noted that 66% of the articles framed sexting as a risky, unhealthy behavior. Many of those articles focused on the relationship between sexting and risky sexual behaviors, but a systematic examination and synthesis of that literature was beyond the scope of her review.
Scholarly interest in the potential links between sexting and risky sexual behavior began with the publication of the results of the Sex and Tech survey (2008) commissioned by the National Campaign to Prevent Teen and Unplanned Pregnancy and Cosmogirl.com. Of the 1,280 teens (ages 13–19) and young adults (ages 20–26) who participated in the study, 39% of teens and 59% of young adults had sent or posted sexually suggestive messages, and 20% of teens and 33% of young adults had sent or posted nude or seminude photos of themselves. Moreover, 38% of teens and 40% of young adults indicated that the act of sexting someone made dating or hooking up with that person more likely. Close on the heels of the Sex and Tech survey came the AP-MTV Digital Abuse Study (2009), which involved a survey completed by a nationally representative sample of individuals ages 14–24. This study documented similar sexting prevalence rates, adding that sexually active individuals were twice as likely to send nude self- images than their nonsexually active counterparts. Findings from these seminal studies, as well as some public sexting scandals, fueled further research on the links between sexting and sexual behavior; however, contradictory findings in the extant literature make it difficult to determine the nature or strength of this relationship (Davis, Powell, Gordon, Kershaw, 2016).
Scholars need a better understanding of the links between sexting and risky sexual behavior to make evidence-based recommendations for policies and programs focused on promoting safer sex and/or digital citizenship. Although this literature is still growing, we have amassed enough studies on sexting and sexual behavior to allow for a meta-analytic review, the chief purpose of which, according Bangert-Drowns (1997), is the “comprehensive, statistical integration of contradictory empirical findings” (p. 244). To this end, we gathered all published, peer-reviewed studies on sexting and sexual behavior and subjected those studies to a meta-analytic and critical review. The following sections describe how researchers have studied sexting and what we can glean from a critical review and synthesis of this literature.
Initially, researchers focused their efforts on establishing the prevalence of sexting within particular populations, but, more recently, scholars turned their attention to sexting correlates (Walrave et al., 2015). Neither set of studies, however, has produced consistent findings. Studies involving nationally representative probability samples of teens report sexting prevalence rates ranging from 2.5% to 24% with an estimated mean of 10.2%. Prevalence also varies across studies involving adult samples, with some reporting rates as low as 30% and others as high as 81% (Klettke, Hallford Mellor, 2014). Scholars (e.g., Ringrose et al., 2012; Strassberg, Rullo, Mackaronis, 2014) have attributed these discrepant findings to various methodological issues, including sampling problems. For example, nonprobability and adult samples frequently yield the highest prevalence rates, and, like most sex research, sexting studies rely heavily on convenience samples (Dunne, 2002; Klettke et al., 2014). The wide variation in sexting prevalence rates might also be due to various biases. As Ringrose and colleagues (2012) noted, “It is difficult to know if ‘sexting’ is under-reported because of social desirability factors (e.g. embarrassment on the part of respondents) or over-reported because of response biases (those who do it may be more likely to respond to surveys).” The lack of a consensus definition or standardized measure of sexting only further complicates matters (Cooper et al., 2016).
These methodological issues might also account for the inconsistent findings regarding the relationship between sexting and risky sexual behavior. For example, some studies have reported a significant relationship between sexting and risky sexual practices, such as sex with multiple partners (e.g., Benotsch, Snipes, Martin, Bull, 2013; Dir, Cyders, Coskunpinar, 2013) or without protection (e.g., Crimmins Seigfried-Spellar, 2014; Yeung, Horyniak, Vella, Hellard, Lim, 2014); others (e.g., Ferguson, 2011; Gordon-Messer, Bauermeister, Grodzinski, Zimmerman, 2013) have found no significant association between sexting and high-risk sexual behavior. There seems to be greater consensus among researchers regarding the relationship between sexting and sexual activity, more generally; however, Temple and Choi (2014), who analyzed the second and third waves of data from a longitudinal sexting study, found that being asked or asking for a sext at Wave 2 was not associated with sexual activity at Wave 3. As such, we sought to clarify the relationship between sexting, sexual activity, and risky sexual practices (i.e., having multiple sex partners or unprotected sex) through a meta-analytic and critical review of this literature.
In addition to these methodological limitations, theoretical issues make it difficult to interpret sexting study findings. Several scholars (e.g., Cooper et al., 2016; Marcum, Higgins, Ricketts, 2014) have decried the dearth of theoretically informed research on sexting and have underscored the need for theory testing and building in this area. Although this meta-analysis was not designed to test a theory, per se, our research questions were informed by two theoretical perspectives, which align with the deviance and normalcy frames that dominate the sexting literature. First, problem behavior theory (Jessor Jessor, 1977) offers a compelling explanation for why sexting and risky sexual behavior might be linked, an assumption of the deviance perspective. Problem behavior theory assumes that different problem behaviors stem from the same causes and that individuals who engage in one problematic activity will be more likely to exhibit other problematic behaviors. The deviance perspective that dominates the sexting literature advances similar arguments regarding sexting and sexual risk. If, in fact, sexting is linked to risky sexual practices, then this might give credence to both the deviance perspective and problem behavior theory.
RQ1: What is the nature of the relationship between sexting and sexual activity?
RQ2: What is the nature of the relationship between sexting and risky sexual practices (i.e., having multiple sex partners or unprotected sex)?
Method
To address our research questions, we relied on meta-analytic techniques, which allow researchers to synthesize the results of multiple studies in order to measure the overall effect of one variable on another. We used a random-effects model, which does not assume a single fixed effect exists between variables in studies carried out in different samples representing different populations (Hedges Vevea, 1998). We followed Lipsey and Wilson's (2001) recommendation to employ Pearson's r as the standard effect size representing the relationships between our concepts of interest: sexting and sexual behaviors. The following sections describe how we approached searching the literature, coding the articles we located, and building a meta- analytic dataset.
Searching the Literature
We began with exhaustive searches using the PubMed and EBSCO databases. In EBSCO, we confined our search to the subdatabases that index social science research, which include Academic Source Complete, Communication Source, ERIC, PsycINFO, and PsycARTICLES. We used one search term, sexting, and we limited our searches to published, English-language articles. We concentrated our efforts on gathering published, peer-reviewed articles for two reasons. First, we wanted to avoid drawing on the same data more than once in our analysis; because many of the conference papers and dissertations that we examined were eventually published, we excluded the unpublished versions. Second, although including unpublished manuscripts in a meta-analysis is one way in which researchers attempt to combat publication bias, Ferguson and Brannick (2012), in an empirical test of this assumption, found that this practice might actually increase publication bias. As such, we focused on published articles on sexting and sexual behavior.
In addition to searching for relevant research via library databases, we reviewed the reference lists of the articles we gathered to identify potentially useful papers that were not indexed by PubMed or EBSCO. Our database searches and reference list mining produced 234 results. After removing duplicates and articles that did not meet our inclusion criteria (i.e., quantitative research on sexting and sexual behavior), we were left with 55 articles.
Coding the Literature
After our literature search, we reviewed each article in order to identify the operational definitions of sexting and sexual behavior that were employed. We identified the following characteristics of each article: the authors; the publication date; the publication outlet; the study design (i.e., cross-sectional or longitudinal); the measures of sexting and sexual behavior; the sample size; and, the characteristics of the sample (i.e., race, age, sex/gender, and sexual identity). After conducting this review of the articles that we had gathered, we decided to exclude 32 articles because their authors either did not report how they measured sexting or did not examine the relationship between sexting and sexual behavior. As a result, we were left with 23 articles that dealt with the variables of interest: sexting, sexual activity, unprotected sex acts, and sex with multiple partners.
Building a Meta-Analytic Data Set
We used a spreadsheet to record the effects representing the relationships between our variables of interest, and we included these effects regardless of whether they were reported in the article as bivariate relationships only or within a multivariate model. For example, in some cases, the variables of interest were only reported in the results of a logistic regression model that also included other independent variables. In line with current recommendations (Ferguson, 2015), we chose to favor results reported from multivariate models, if available, or bivariate relationships (e.g., correlation, chi-square with one degree of freedom). Of the 23 articles that dealt with our variables of interest, eight did not have sufficient information to use for our meta-analysis (i.e., two used analytic categories for sexting that were not comparable to those in other articles, three measured relevant variables but did not report any relationships between them, and three used variables that measured more than simply sexting or sexual behavior and were not comparable to other articles). Due to differences in the way in which these 23 reported articles data, we contacted the corresponding authors in order to obtain the necessary information, but our requests were met with few favorable responses. Unable to secure the necessary data to conduct our own multivariate analyses, we were faced with a choice between: (a) a meta-analytic data set that was incomplete due to nonresponse from some authors or (b) a data set that contained estimates derived solely from information provided in the articles, thereby excluding the eight with insufficient information. Either choice seemed imperfect, but, ultimately, we chose the second option.
The 15 articles with usable data included six that reported multivariate models (logistic regression and path analysis/multiple regression) using sexting as an independent variable and sexual behavior as a dependent variable. We were able to use effect estimates taken directly from these articles. Three articles reported only bivariate correlations. Six other articles reported multivariate results with sexual behavior as an independent variable and sexting as a dependent variable, which ran counter to the assumptions of the theories that guided this analysis and produced effect estimates that were not analytically comparable to those from the other articles. Thus, for these studies, we calculated the effects as bivariate relationships (chi-square tests with one degree of freedom) from information available in the articles. We were able to maintain a high level of quality control by converting the data, as reported in the articles, using a consistent set of formulas provided in the methodological literature on meta-analysis (Lipsey Wilson, 2001). Our final meta-analytic spreadsheet included 15 articles drawing on 14 unique data sets (two articles were based on the same data set). Study characteristics are summarized in Table 1.
Table 1. Characteristics of Included Studies Click for the table https://academic.oup.com/view-large/96732588
These 15 articles identified a total of 50 effects. Eleven effects were originally reported in the articles as a bivariate Pearson's r correlation or equivalent (e.g., phi coefficient, standardized regression coefficient), requiring no mathematical conversion. The remaining 39 effects required a mathematical conversion to Pearson's r. One effect was reported as a comparison of means between two groups (i.e., individuals who had sexted and individuals who had not sexted), for which we calculated the standardized mean difference and converted it to a correlation coefficient. Ten effects were reported as adjusted odds-ratios. The remaining 28 effects were either (a) reported as chi-square tests with one degree of freedom or (b) derived from information provided in the article and used to calculate a chi-square test ourselves. For the latter, we used published frequency counts or percentage breakdowns to construct cross-tabulation tables between sexting (reported as a “yes” or “no” dichotomy in each article) and sexual behaviors. We could then conduct significance tests for a chi-square statistic with one degree of freedom. We employed conversion formulas for effect sizes from Lipsey and Wilson (2001), SPSS macros from Wilson (2005), and the formula for I2 from Higgins and Thompson (2002). An overview of the effects we identified from each article is provided in Table 2. Click for the table 2 https://academic.oup.com/view-large/96732590
Results
Our analysis relied on data derived from a total of 15 published articles that explored possible links between sexting and aspects of sexual behavior, including participants' engagement in sexual acts generally, engagement in unprotected sexual acts, and number of sexual partners. In the following sections, we expound on the characteristics of these studies and present the results of our meta-analytic review.
Review of Study Characteristics
Article details
Our critical review of the articles included in our meta-analysis uncovered some trends with respect to the publication of manuscripts that dealt with sexting and sexual behavior as well as the designs of these kinds of studies. Although we did not restrict our literature search by date, we did not locate any articles on sexting and sexual behavior that were published prior to 2011. In addition, a majority (53.3%) of the studies were published in the last 2 years (i.e., 2014–2015). We also noticed some trends regarding publication outlets. The vast majority (73.3%) of these articles appeared in health journals (see Table 1). Conspicuously absent from this list of publication outlets were communication journals, despite the centrality of communication and communication technologies to sexting.
A closer reading of the included articles revealed additional trends regarding study designs. For example, despite the lack of prior research on which to base hypotheses, none of the authors posed formal research questions. Instead, authors specified the general aims or purpose of the study. Six articles described the authors' hypotheses, and, although these hypothesis statements made multiple predictions, only one article (i.e., Dir et al., 2013) presented enumerated hypotheses. Only two of the articles that we reviewed based their predictions on a specific theory or theoretical model (i.e., Dir Cyders, 2015; Dir et al., 2013), and all but one study (i.e., Temple Choi, 2014) relied on cross-sectional data.
All but one study (i.e., Dake et al., 2012) detailed the specific items or instruments used to measure sexting, and these measures implicated different behaviors, technologies, content, and contexts. The majority (85.7%) of the 14 studies that described their sexting measures emphasized “sending” in their operational definitions; seven of these studies also assessed “receiving” sexts. One study (i.e., Jonsson et al., 2015) included a question about “posting” sexual content, and Ybarra and Mitchell (2014) queried participants about having “sent or shown” sexual images of oneself. The communication technologies associated with sexting also differed across studies. In seven studies, participants were asked about sexting via text or online, and five studies focused just on text messaging. The measure used by Ferguson (2011) did not reference a particular medium or technology, and Ybarra and Mitchell's (2014) operational definition included sharing sexual pictures “in person, on the Internet, and on cell phones and text messaging” (p. 758). The measures employed in these studies also characterized the form and content of sexts in different ways. Eight studies focused entirely on images, including pictures and videos; the remaining studies (n = 6) assessed both images and text. Measures most commonly referred to sext content as sexually explicit, suggestive, or provocative in nature (n = 7). Other measures simply referred to sext content as “sexual” in nature (n = 2), and six studies included questions regarding nude, nearly nude, or seminude photos. Ten studies specified the photo or message subject (i.e., images of or messages about oneself), but only Yeung and colleagues (2014) asked about the relational context in which sexting occurred.
Sample details
The articles we reviewed reported data from a total of 18,190 participants, and their characteristics are reported in Table 1. Sample sizes ranged from 88 to 3,715 (M = 1,212.7). Eleven articles described samples with a higher percentage of female than male participants, and, in the studies with more male participants, the percentage of males who participated was only slightly higher than the percentage of females who took part. Of the four studies that mentioned transgender participants (i.e., Jonsson et al., 2015; Rice et al., 2012, 2014; Ybarra Mitchell, 2014), data collected from these participants was excluded from the analysis in all but the 2012 publication from Rice and colleagues. One study (i.e., Houck et al., 2014) did not describe the demographic data, including the sex/gender or racial composition, for the sample as a whole; instead, the authors included data for subsamples of participants. Ybarra and Mitchell (2014) used a similar technique when reporting the racial composition of their sample. Two articles (i.e., Jonsson et al., 2015; Yeung et al., 2014) did not specify the race or ethnicity of those who participated. Six articles described samples dominated by White participants, four indicated that the samples featured more Hispanic individuals than people from other ethnic groups, and one study sample was comprised of more African Americans than Whites or Hispanics. Six articles mentioned questions regarding participants' sexual identity or orientation (i.e., Dir Cyders, 2015; Gordon-Messer et al., 2013; Rice et al., 2012, 2014; Ybarra Mitchell, 2014; Yeung et al., 2014); the remaining studies either did not include a sexual identity measure or did not report this data. Finally, although the participants in these studies ranged in age from 10 to 51 years old, only one study (i.e., Dir Cyders, 2015) involved individuals over 30 years old, and, given that the mean age of participants in that study was 21.2 years, those individuals appeared to be outliers. Four of the studies were mainly comprised of adolescents (typically defined as ages 12–18); four others consisted of emerging adults (i.e., individuals between the ages of 18 and 25); and, six studies did not report the age range of the participants. None of the articles that we reviewed focused entirely on adult samples (i.e., individuals 26 or older).
Meta-Analytic Review
Research Question 1 concerned the nature of the relationship between sexting behavior and engagement in sexual activity in general.1 A total of 10 studies provided data relevant to this research question, with 23 effects reported (see Table 3). After averaging effects for each study with more than one effect reported, we conducted a random-effects model. The results produced a medium effect size (Cohen, 1988) of r = .35 (95% confidence interval between .23 and .46). Based on these results, the 10 studies provide evidence for the assertion that sexting is associated with engaging in sexual activity. There is also evidence of a great deal of heterogeneity in the effect size across studies based on the coefficient I2, where a value closer to 100% indicates more heterogeneity (Higgins Thompson, 2002). Specifically, the effects for studies using this dependent variable reported relationships anywhere from small (r = .10) to large (r = .66; Cohen, 1988).
Results
Our analysis relied on data derived from a total of 15 published articles that explored possible links between sexting and aspects of sexual behavior, including participants' engagement in sexual acts generally, engagement in unprotected sexual acts, and number of sexual partners. In the following sections, we expound on the characteristics of these studies and present the results of our meta-analytic review.
Review of Study Characteristics
Article details
Our critical review of the articles included in our meta-analysis uncovered some trends with respect to the publication of manuscripts that dealt with sexting and sexual behavior as well as the designs of these kinds of studies. Although we did not restrict our literature search by date, we did not locate any articles on sexting and sexual behavior that were published prior to 2011. In addition, a majority (53.3%) of the studies were published in the last 2 years (i.e., 2014–2015). We also noticed some trends regarding publication outlets. The vast majority (73.3%) of these articles appeared in health journals (see Table 1). Conspicuously absent from this list of publication outlets were communication journals, despite the centrality of communication and communication technologies to sexting.
A closer reading of the included articles revealed additional trends regarding study designs. For example, despite the lack of prior research on which to base hypotheses, none of the authors posed formal research questions. Instead, authors specified the general aims or purpose of the study. Six articles described the authors' hypotheses, and, although these hypothesis statements made multiple predictions, only one article (i.e., Dir et al., 2013) presented enumerated hypotheses. Only two of the articles that we reviewed based their predictions on a specific theory or theoretical model (i.e., Dir Cyders, 2015; Dir et al., 2013), and all but one study (i.e., Temple Choi, 2014) relied on cross-sectional data.
All but one study (i.e., Dake et al., 2012) detailed the specific items or instruments used to measure sexting, and these measures implicated different behaviors, technologies, content, and contexts. The majority (85.7%) of the 14 studies that described their sexting measures emphasized “sending” in their operational definitions; seven of these studies also assessed “receiving” sexts. One study (i.e., Jonsson et al., 2015) included a question about “posting” sexual content, and Ybarra and Mitchell (2014) queried participants about having “sent or shown” sexual images of oneself. The communication technologies associated with sexting also differed across studies. In seven studies, participants were asked about sexting via text or online, and five studies focused just on text messaging. The measure used by Ferguson (2011) did not reference a particular medium or technology, and Ybarra and Mitchell's (2014) operational definition included sharing sexual pictures “in person, on the Internet, and on cell phones and text messaging” (p. 758). The measures employed in these studies also characterized the form and content of sexts in different ways. Eight studies focused entirely on images, including pictures and videos; the remaining studies (n = 6) assessed both images and text. Measures most commonly referred to sext content as sexually explicit, suggestive, or provocative in nature (n = 7). Other measures simply referred to sext content as “sexual” in nature (n = 2), and six studies included questions regarding nude, nearly nude, or seminude photos. Ten studies specified the photo or message subject (i.e., images of or messages about oneself), but only Yeung and colleagues (2014) asked about the relational context in which sexting occurred.
Sample details
The articles we reviewed reported data from a total of 18,190 participants, and their characteristics are reported in Table 1. Sample sizes ranged from 88 to 3,715 (M = 1,212.7). Eleven articles described samples with a higher percentage of female than male participants, and, in the studies with more male participants, the percentage of males who participated was only slightly higher than the percentage of females who took part. Of the four studies that mentioned transgender participants (i.e., Jonsson et al., 2015; Rice et al., 2012, 2014; Ybarra Mitchell, 2014), data collected from these participants was excluded from the analysis in all but the 2012 publication from Rice and colleagues. One study (i.e., Houck et al., 2014) did not describe the demographic data, including the sex/gender or racial composition, for the sample as a whole; instead, the authors included data for subsamples of participants. Ybarra and Mitchell (2014) used a similar technique when reporting the racial composition of their sample. Two articles (i.e., Jonsson et al., 2015; Yeung et al., 2014) did not specify the race or ethnicity of those who participated. Six articles described samples dominated by White participants, four indicated that the samples featured more Hispanic individuals than people from other ethnic groups, and one study sample was comprised of more African Americans than Whites or Hispanics. Six articles mentioned questions regarding participants' sexual identity or orientation (i.e., Dir Cyders, 2015; Gordon-Messer et al., 2013; Rice et al., 2012, 2014; Ybarra Mitchell, 2014; Yeung et al., 2014); the remaining studies either did not include a sexual identity measure or did not report this data. Finally, although the participants in these studies ranged in age from 10 to 51 years old, only one study (i.e., Dir Cyders, 2015) involved individuals over 30 years old, and, given that the mean age of participants in that study was 21.2 years, those individuals appeared to be outliers. Four of the studies were mainly comprised of adolescents (typically defined as ages 12–18); four others consisted of emerging adults (i.e., individuals between the ages of 18 and 25); and, six studies did not report the age range of the participants. None of the articles that we reviewed focused entirely on adult samples (i.e., individuals 26 or older).
Meta-Analytic Review
Research Question 1 concerned the nature of the relationship between sexting behavior and engagement in sexual activity in general.1 A total of 10 studies provided data relevant to this research question, with 23 effects reported (see Table 3). After averaging effects for each study with more than one effect reported, we conducted a random-effects model. The results produced a medium effect size (Cohen, 1988) of r = .35 (95% confidence interval between .23 and .46). Based on these results, the 10 studies provide evidence for the assertion that sexting is associated with engaging in sexual activity. There is also evidence of a great deal of heterogeneity in the effect size across studies based on the coefficient I2, where a value closer to 100% indicates more heterogeneity (Higgins Thompson, 2002). Specifically, the effects for studies using this dependent variable reported relationships anywhere from small (r = .10) to large (r = .66; Cohen, 1988).
Table 3 :Meta-Analytic Results Regarding the Association between Sexting and Sexual Behaviors. Click for the table 3 https://academic.oup.com/view-large/96732602
With Research Question 2, we sought to clarify the relationship between sexting behavior and unprotected sexual activity. For these variables, nine studies provided a total of 14 effects (see Table 3); for studies reporting multiple effects, we calculated an average effect size with each effect weighted equally. The random-effects model reported an effect size of r = .16 (95% confidence interval between .09 and .23). The result suggests that sexting co-occurs with unprotected sexual behavior, although the relationship is small. As with the previous dependent variable, the coefficient I2 indicates a high level of heterogeneity across the effects included in this analysis (ranging from a small effect of r = .00 to a medium effect of r = .36; Cohen, 1988).
RQ2 also focused on the relationship between participants' sexting behavior and number of sexual partners. There were seven studies reporting 13 effects for this relationship (see Table 3). Multiple effects within an individual study were averaged. The random-effects results indicated an effect size of r = .20 (95% confidence interval between .16 and .23). This result suggests a consistent but small association (Cohen, 1988) between engaging in sexting behavior and having a higher number of sexual partners. Contrary to the other two dependent variables, the coefficient I2 suggests more consistency (i.e., less heterogeneity) within this group of effects, which ranged from small (r = .10) to medium (r = .32).2
Discussion
Sexting and its potential links to sexual behavior have garnered significant scholarly scrutiny in recent years, but this literature is marked by divergent perspectives and disparate findings (Doring, 2014). The deviance perspective that dominates this literature assumes that sexting is a risky practice and, therefore, should be associated with other risky practices, such as having multiple sex partners or unprotected sex; however, an emerging normalcy discourse challenges these assumptions (Cooper et al., 2016). To assess these claims, we conducted a meta-analytic and critical review of published research on sexting and its relationship to sexual behavior. Our review revealed several trends in the sexting literature, including the types of study designs and samples that were most common. In addition, we uncovered considerable variability in how sexting was measured across the 15 studies including in our analysis. Despite this variability, our meta-analysis was able to parse out significant relationships between sexting and three aspects of sexual behavior: general sexual activity, history of unprotected sex, and number of sexual partners. The following sections describe possible explanations for and implications of each of these findings.
First, our meta-analytic review of the 10 studies and 23 effects that dealt with sexting and general sexual activity revealed a substantive relationship between two, which undermines the catharsis argument. Although this finding might confirm parental concerns about sexting, these results must be interpreted with caution because the vast majority of studies reviewed were cross-sectional in nature. As such, all we know is that sexting and sexual behavior co-occur. Sexting might act as a gateway......
Indrasish Roy 2 yrs
Very well documented