Abstract
This empirical study investigates adolescents’ responses to peers’ personal accounts of romantic relationship difficulties posted to an online forum. We analyze 440 anonymous responses to personal accounts of four romantic relationship issues: controlling partners, break-ups, trust issues, and partner cruelty. Responses were categorized, in order of prevalence, as judgment, recommendation, words of kindness, personal experience, and probing question. Adolescents who provide recommendations most often advise that peers terminate their relationships, rather than seek help or communicate directly. However, adolescents respond differently to cruelty, as compared with the other three relationship issues. Although still likely to recommend ending the relationship in cases of cruelty, they are significantly more likely to suggest seeking help, and less likely to suggest direct communication. Adolescents also respond differently depending on the poster’s gender: They offer more Recommendations overall, and specifically more Recommendations to Seek Help and Break-up/Move on, to female than to male posters.
For adolescents who are beginning to experience and form romantic connections, few topics are more salient on a daily basis than their romantic relationships (Carver, Joyner, & Urdy, 2003; Furman & Hand, 2006). A recent national survey found that fully two thirds of middle and high school students were either currently in or had recently been in romantic relationships (Zweig, Dank, Lachman, & Yahner, 2013). Healthy romantic relationships during adolescence serve as a source of positive affect, bolster feelings of self-worth and competence, and form the basis for positive relationships in adulthood (see Collins, Welsh, & Furman, 2009, for discussion).
However, difficulties that arise in adolescent romantic relationships can also be detrimental. In the short term, romantic relationship difficulties can result in psychosocial distress (Joyner & Udry, 2000; McIsaac, Connolly, McKenney, Pepler, & Craig, 2008), behavioral problems, and poor academic performance (Zimmer-Gembeck, Siebenbruner, & Collins, 2001). The consequences intensify for adolescents in abusive relationships, which are associated with issues such as unhealthy weight control, substance abuse, risky sexual behavior, and suicide ideation (Silverman, Raj, Mucci, & Hathaway, 2001). In the long term, destructive adolescent relationship skills and patterns can also have lasting effects on adult relationships (Barber & Eccles, 2003).
Considering the substantial influence of romantic relationships on adolescents’ psychosocial and behavioral functioning, it is crucial that adolescents successfully navigate romantic relationship issues when they arise and, relatedly, that they receive useful, high quality advice when they seek it from others. For their romantic relationship queries, adolescents overwhelmingly turn to peers rather than to adults or professionals (Black, Tolman, Callahan, Saunders, & Weisz, 2008; Fry et al., 2014; Tishby et al., 2001; Watson, Cascardi, Avery-Leaf, & O’Leary, 2001). With unprecedented access to the Internet, adolescents now increasingly seek peer advice on myriad romantic relationship issues through online platforms (Weinstein & Selman, 2014; Gould, Munfakh, Lubell, Kleinman, & Parker, 2002; Suzuki & Calzo, 2004).
Yet, research has not systematically explored the content of the support adolescents provide online to peers experiencing romantic relationship difficulties. We address this gap in the current study, in which we examine adolescents’ responses to peers’ personal accounts of four types romantic relationship difficulties posted to a popular online forum. To our knowledge, this is the first study to examine, exclusively and in some depth, the content of anonymous multi-user online peer support for different kinds of adolescent romantic relationship difficulties. Our findings illuminate practical implications for how to improve and complement online peer support to amplify its positive impact.
Types of Romantic Relationship Difficulties and Their Prevalence
Physical and Psychological Abuse
A large body of literature, dating back to at least the 1980s, examines physical abuse in adolescent romantic relationships (see Lewis & Fremouw, 2001, for discussion). Depending on definitional criteria and the demographics of the sample, these studies find that between 9% and 46% of adolescents experience physical abuse in relationships (Feiring, Deblinger, Hoch-Espada, & Haworth, 2002; Grunbaum et al., 2002; Jouriles, Garrido, Rosenfield, & McDonald, 2009; Watson et al., 2001).
Although it is more common than physical abuse (Feiring et al., 2002; Jackson, Cram, & Seymour, 2000), psychological abuse in relationships receives less empirical attention. Psychological abuse is often used as a blanket term for several forms of abuse, and documented prevalence rates vary substantially among studies. One study found that 59% of youth experience psychological abuse when the criteria included insult, ridicule, hostile voice, blaming, threatening break-up, and silent treatment; the prevalence rate rises to 91% when abuse criteria also includes blame, arousing feelings of jealousy, guilt, anger, accusation for flirting with another person, and controlling behavior (Jouriles et al., 2009). When psychological abuse is defined as including “monopolization” (e.g., “keeps tabs on where I go, who I am with”), “degradation” (e.g., “slams down my opinions”), and “isolation” (“ignores me as if I don’t exist”), 81.5% of female and 76.3% male high school students experienced such abuse (Jackson et al., 2000). Psychological abuse is also quite distressing: Adolescents at times perceive psychological abuse as more unpleasant and intentionally hurtful than physical aggression (Jouriles et al., 2009). Moreover, psychological aggression is one of the best predictors of physical aggression (O’Leary & Smith Slep, 2003). Thus, while physical abuse is more commonly explored, the inclusion of psychological abuse in studies of adolescents’ romantic struggles is crucial.
In general, studies report inconsistent findings regarding gender differences in dating violence perpetration and victimization (see Hickman, Jaycox, & Aronoff, 2004, for discussion). Although some studies find that boys actually report a higher level of victimization and lower perpetration of physical and psychological aggression (Spencer & Bryant, 2000; Swahn, Simon, Arias, & Bossarte, 2008), others find no gender difference (O’Keefe, 1997), and higher rates of victimization among girls (Watson et al., 2001).
Breaking-Up
The issues with which adolescents grapple in the context of romantic relationships are by no means limited to abuse. Experiencing the dissolution of a romantic relationship—“breaking-up”—is a common issue that can be challenging for adolescents, eliciting psychological distress, behavioral difficulties, and diminished academic performance (Field, Diego, Pelaez, Deeds, & Delgado, 2009). A large proportion of adolescents experience break-ups: By the time students are in college, more than 90% have rejected or been rejected by a romantic partner (Baumeister, Wotman, & Stillwell, 1993). The process of breaking-up is often complicated by ongoing contact with and unwanted pursuits by a previous romantic partner (Langhinrichsen-Rohling, Palarea, Cohen, & Rohling, 2000). Break-ups are distressing for both males and females: Although some studies find that girls experience more post-break-up distress (Fisher, 2004; Mearns, 1991), others find that boys experience more distress (Knox, Zusman, Kaluzny, & Cooper, 2000).
Peer Support for Adolescent Romantic Relationship Difficulties
Adolescents’ Preference for Peer Support
When adolescents face issues in their romantic relationships, whether related to partner cruelty, controlling tendencies, or relationship dissolution, where do they turn? The literature on adolescent help-seeking behavior consistently highlights the essential position of adolescents in providing support for their peers’ romantic relationship difficulties (Black et al., 2008; Fry et al., 2014; Watson et al., 2001). Rarely do adolescents report seeking help from parents, or from school or health professionals, law enforcement, or community resources for romantic relationship issues (Molidor & Tolman, 1998; Ocampo, Shelley, & Jaycox, 2007) In fact, many adolescents actually do not seek help at all for romantic relationship difficulties (Ashley & Foshee, 2005; Black et al., 2008; Fry et al., 2014; Jackson et al., 2000).
Some barriers to help-seeking identified in previous research include embarrassment, fear of negative judgment from others, fear of pressure from others to break-up, concerns about confidentiality and independence, and lack of knowledge of where to go for professional services (Christiana et al., 2000; Gallopin & Leigh, 2009; Mahlstedt & Keeny, 1993; Sears, Byers, Whelan, & Saint-Pierre, 2006). The potential for both autonomy and anonymity when adolescents seek help online can mitigate concerns that interfere with adolescents’ offline help-seeking for romantic issues (McKenna & Bargh, 2000; Suler, 2004).
Adolescent boys might be in a particularly vulnerable position as they are less likely than girls to seek help when experiencing distress (Rickwood, Deane, Wilson, & Ciarrochi, 2005; Tishby et al., 2001). This gender difference in help-seeking behavior could ostensibly stem from cultural values that define masculinity in relation to self-reliance, physical toughness, and emotional control (Addis & Mahalik, 2003). Such gender expectations might serve as obstacles for adolescent boys in admitting to a personal difficulty and relying on others for help and predispose them to accept abusive relationships (Jackson et al., 2000). The potential to access peer support through the Internet might therefore be especially valuable for adolescent boys, offering them a more anonymous context in which to share struggles and seek support. Indeed, Gould and her colleagues (2002) found that although boys are less likely than girls to seek help offline, they are just as likely as girls to seek help on the Internet.
Adolescents’ Responses to Peers’ Romantic Relationship Difficulties
Despite the pervasiveness of adolescent relationship difficulties and the tendency among adolescents to rely on peers, few studies have examined exactly how adolescents serve as advisors for peers’ romantic relationship difficulties offline, especially when the challenges involve issues other than physical abuse. In addition, much of the research is based on responses to hypothetical scenarios (Ocampo et al., 2007), self-reports by victims describing the peer support they previously received (Fry et al., 2014; Weisz, Tolman, Callahan, Saunders, & Black, 2007), or self-reports by adolescents who provided peer support in the past (Fry et al., 2014).
These studies indicate that adolescents report a variety of strategies to help peers with romantic relationship difficulties. For friends experiencing dating violence, adolescents respond by telling the friend to leave the partner (80%), talking with the friend (79%), taking direct action (76%), telling the friend to talk to adult (50%), and telling the friend to call a hotline (19%; Fry et al., 2014). However, advice to leave the partner often lacks concrete steps for how to actually leave the abuser or efforts to connect the victim with additional helpers (Ocampo et al., 2007). In addition, Weisz et al. (2007) reported that the more severe a dating violence incident is, the more avoidant peers are in their responses; those who are most in need of help may therefore be least likely to receive peer advice and support.
Peer Support in Cyberspace
Earlier studies indicate that romantic relationship issues are among the most prevalent topics of inquiry on Internet forums for adolescents (Weinstein & Selman, 2014; Suzuki & Calzo, 2004); yet, little is known about the responses adolescents receive when they solicit online support for these issues. Although Gould et al. found in 2002 that most adolescents report low satisfaction levels about the help they receive for emotional problems on the Internet, there is a dearth of research that illuminates why adolescents might be dissatisfied with the help. An exception is Suzuki and Calzo’s (2004) investigation of peer advice on moderated online bulletin boards for adolescents. They found that adolescents provided advice in the forms of personal opinions (63%), actionable advice (44%), concrete information (37%), personal experiences (33%), and emotionally supportive comments (12%) to peers’ posts about romantic relationships, physical health, body image, and sex. However, this study (and Gould et al.’s, 2002, study) was conducted over a decade ago, and it did not focus in-depth on advice for different kinds of romantic relationship issues; the content of online peer feedback merits an updated and more detailed examination.
Research Questions
The current study is an examination of adolescents’ anonymous responses to peers’ romantic relationship difficulties posted on an online forum. We explore three questions as follows:
Method
Demographics of Users on the ATL (A Thin Line)
The stories and comments that we explore in the current study were posted by adolescents to MTV’s ATL platform. The U.S.-based company launched ATL (http://www.athinline.org) in 2010 as an initiative to support adolescents struggling with digital issues. The website provides a platform dedicated to users sharing their personal experiences and exchanging feedback. Posters of the ATL website have the option of identifying their name, gender, and age when posting personal accounts. 65% of the posters of personal accounts reported their gender (86.3% female, and 13.7% male), and 61.8% of the posters reported their age (Mage = 16.3, median and modal age = 15). Among posters who reported their age, less than 1% report they are 12 or under; 51% are 13-15; 34% are 16-18; 8% are 19-21; and 5% are 22 or over. Responders to the personal accounts, however, remain anonymous and do not have the option of providing age and gender. Although not every user of the ATL is or reports to be an adolescent, a majority of participants on the site are between the age of 13 and 18. We, therefore, consider the peer responses as heavily representative of adolescent perspectives and use the term adolescent to describe the sample.
Sampling Criteria and Types of Expressed Romantic Relationship Difficulties
The data set that we accessed through our licensing agreement with MTV included all personal accounts and responses posted to the site between March 2010 and July 2013: 7,146 personal accounts and 24,409 response comments.
In a previous analysis of 2,000 stories systematically sampled from the current data set, Weinstein and Selman (2014) found that, while the site was originally intended for digital issues, adolescents used it predominantly for issues related to romantic relationships. The authors found four reliably identifiable categories of romantic relationship difficulties expressed in adolescents’ personal accounts: CONTROLLING BEHAVIOR (96 personal accounts, 416 responses), BREAK-UPS (198 personal accounts, 515 responses), CRUELTY (36 personal accounts, 137 responses), and TRUST ISSUES (298 personal accounts, 1,004 responses; the names—but not the categories—are modified slightly in the current study). These categories are not mutually exclusive and though the poster is most often the victim of the reported relational transgression, some posters are self-described perpetrators or witnesses.
CONTROLLING BEHAVIOR captures accounts that describe difficulties involving a partner who discourages the poster from spending time with others or going places without the partner; engages in constant contact to “keep tabs” on whereabouts and activities; exerts pressure to share personal email, social media accounts, or digital correspondence with others; and/or secretly trail the partner. BREAK-UPS includes accounts on a range of issues related to relationship termination, from distress or uncertainty over a recent break-up to difficulties in dealing with an ex-partner following a break-up. Personal accounts about CRUELTY include both psychological and physical abuse of a partner, such as name-calling, swearing, put-downs, threats, yelling, or aggressive displays of anger. Almost none of the stories in the data set explicitly described acts of physical abuse; however, several posters described partners or certain behavior as “abusive,” potentially an implicit indication of physical abuse. The TRUST ISSUES category includes both accounts about posters cheating or being cheated on by their partners, and/or feelings of insecurity, such as jealousy and distrust. For the current study, we organized personal accounts and their corresponding responses by issue type (i.e., CONTROLLING BEHAVIOR, BREAK-UPS, CRUELTY, TRUST ISSUES). Although the numbers of responses vary for the four romantic relationship difficulties, for our analysis we randomly selected a total of 110 responses posted to stories in each of the four categories of romantic difficulties, resulting in a total of 440 responses.
The personal accounts in the four types of romantic relationship difficulties elicit different numbers of peer responses: As indicated by the mean number of responses that personal accounts received for each of the four categories of romantic relationship difficulties, adolescent responders are most likely to respond to CONTROLLING BEHAVIOR (average n responses per account = 4.33), followed by CRUELTY (n = 3.81) and TRUST ISSUES (n = 3.36), while they are least likely to respond to BREAK-UPS (n = 2.69).
Development of Response Codes
To capture the kinds of peer advice provided for adolescents who express romantic relationship difficulties, we coded our data using an emic, qualitative content analysis process. Coding is “a typical categorizing strategy in qualitative research” (Maxwell, 2012, p. 106), and researchers use an array of strategies to the process of identifying, developing, and applying codes. Our goal was to inductively develop substantive, descriptive categories that would enable us to explore cross-cutting themes, while using codes that “stay close to the data categorized” (Maxwell, 2012, p. 108); in this sense, our approach is similar to what Hsieh and Shannon (2005) term a “conventional” content analysis.
We began with several word-by-word readings of the posts (Miles & Huberman, 1994). Through these “close” readings, we were purposeful in our effort to familiarize ourselves with the full data set. We then selected a sub-set of stories for a version of “open coding” (e.g., Corbin & Strauss, 2014). Rather than merely labeling words, our open coding process involved searching for substantive categories and preparing memos as we noted recurring recommendations and identified emerging codes (Maxwell, 2012). A team discussion of these memos generated a preliminary list of 13 response codes. We drew on the concept of comparing and contrasting, asking, “how are they alike, and how are they different?” (Smith, 1979, p. 338), as we defined, combined, and refined code categories. We next conducted two rounds of independent coding, each with two coders and each followed by discussion of coding differences; we used a qualitative coding software, NVivo 9, to conduct inter-rater reliability checks. Through these two rounds of coding and discussion, we refined and merged the 13 codes to arrive at the final set of 9 major codes and our corresponding codebook. A subsequent round of reliability testing confirmed sufficient inter-rater reliability for each of the 9 codes (kappas ranging from .74-1.0; Mayring, 2000). The 9 codes, their definitional criteria, and corresponding examples are summarized in Table 1.
Definitions and Examples of Response Codes.
We then coded each response for one or more of the applicable codes, depending on its content. For example, we assigned three codes to the following response:
Your boyfriend sounds very controlling. He isn’t fair to you, and he obviously has trust issues. (Judgment) You should consider whether or not you’re willing to be with such a controlling person. (Recommendation: Other) Talk to him about why he doesn’t trust you. (Recommendation: Communicate)
We totaled prevalence rates and calculated 95% confidence intervals for each response code for the sample as a whole and for each of the four romantic relationship difficulties, which we present in the first section of the findings. Although our content analysis enables us to explore these frequencies and variances, the approach is especially valuable because the categorization process also facilitates our ability to surface and explore variations within response codes, which we also describe. Finally, we present results of a series of chi-square tests that we conducted to examine overall differences in the responses offered for the four romantic relationship difficulties.
Findings
Prevalence of Response Codes
Table 2 shows the total proportions of response codes in the current sample of 440 responses and their 95% confidence intervals. The most popular type of response provided on ATL in response to romantic relationship issues is Judgment, present in 69% (n = 304) of the 440 responses in the entire sample. Recommendation, the second most common type of feedback, is present in 66% (n = 291) of the responses. Statistically, responders are no more likely to provide Judgment than Recommendation, and vice versa. However, responders are significantly more likely to provide Recommendation and Judgment than Words of Kindness (9%, n = 41), Personal Experience (9%, n = 41), and Probing Question (5%, n = 23). Among the sub-codes for Recommendation, responders are significantly more likely to suggest peers Break-Up/Move on (36%, n = 156) than Seek Help (10%, n = 44) or Communicate (14%, n = 62).
Total Proportions of Codes.
Note. Confidence intervals indicated in the brackets are calculated at .05 significance level.
Content Analysis of Different Types of Responses
Judgment
Personal judgments and reflections about the situation are the most common type of feedback offered by adolescents, present in 69% of the responses. Several notable patterns emerge among Judgment responses. First, many responders’ judgments attribute permanence to the problematic behavior or character of the perpetrator, and warn the posters that the situation will not improve (e.g., “This is a cycle that usually doesn’t end”). Others go a step further and predict negative consequences or exacerbation of the problem (e.g., “eventually he will be so controlling he will beat you and you do not need that”). Some responders attempt to provide deeper insight into the situation by elaborating the perspective of perpetrator (e.g., “he’s tellin u these things becuz he’s too scared to face the fact tht maybe HE’S the one who can never get sumone or he has a low self esteem so he takes it out on u”). Judgments often take the form of pure name-calling or strong disapproval of the perpetrator (e.g., “He is a real a******,” “STALKER! CREEPER! CRAZY!”).
Recommendation
The majority of responses on the platform (66%) include at least one Recommendation, which we define as a suggestion characterized by actionable advice. We found three recurring types of Recommendation: Break-Up/Move on, Communicate directly with the partner, and Seek Help.
Break-Up/Move on
Recommendations that advise the poster to simply “dump,” “leave,” or “break-up” with the partner are present in 36% of responses. Some suggestions indirectly recommend breaking-up with the partner by advocating that the poster find a different romantic partner (e.g., “woahh that sucks find someone else”). For posters who express difficulties with an ex-partner, responders recommend that the poster simply “move on,” ignore the ex-partner (e.g., “stop messaging him back”), or take a more extreme measure of blocking the ex-partner. However, none of the responses elaborate on how to carry out the break-up or manage the emotional aftermath for either party.
Communicate
Responses that advise communication with the partner comprise 14% of the responses, ranging from general suggestions (e.g., “talk to him”) to explicit delineations of what the poster should say (e.g., “just tell him you want him to work on being less controlling and say u need a break from him”). Some responders suggest communication with the partner to reassure the partner of the relationship’s strength (e.g., “Assure him that he’s your one and only”), while others suggest confronting the partner to point out problematic behavior (e.g., “tell him straight to face that the relationship is not fair”). The recommendation to communicate is also suggested as a strategy to learn more about the partner’s feelings and thoughts that might underlie the problem (e.g., “You and that other person sound like they have different views on relationships. Talk to him and find out”). When the poster is the perpetrator of the problem in the relationship, some responders suggest communication to directly apologize to the partner.
Seek Help
Recommendations to seek help from a third party are present in 10% of the responses. Of the 44 responses that suggest Seek Help, 19 suggest help from parents and/or adults; 11 responses identify the “police” or “cops,” and 8 advise that the poster “report” or press charges against the significant other; 5 responses provide general recommendations to “get help” or to tell “someone.” School officials, mental health professionals, or peers are not sources of help that adolescents recommend; only one response explicitly mentions “counseling,” and none mention talking to a friend or school official.
Other Recommendation
Recommendations coded as Other, present in 21% of the responses, encompass any advice on what to do, what to think, or how to feel that is not captured by the Seek Help, Break-Up/Move on, or Communicate categories. Responses in this category vary considerably without any clear overarching themes. Some responses are general (e.g., “live life,” “be careful”), while others are explicit about what actions to take (e.g., “Don’t reward his bad behavior with a response. Tell him you will ignore him for an hour the next text you get like that. And for 2 hours after the NEXT one, and 3 hours after THAT . . . . and stick to it”). Some responders offer advice on how the poster should change his or her attitude or approach about the problem (e.g., “just don’t let it bother you”). For posters who are struggling with an ex-boyfriend, several responders recommend directing focus on other people or activities (e.g., “try to get back in the crowd. hang out with old friends or even new friends”).
Words of Kindness
Words of Kindness are present in 9% of the responses. Adolescent responders express kindness and empathy in several ways. Some indicate that they feel really “bad” or “sorry” for the poster while others provide encouragement (e.g., “keep your head up). Kindness is also expressed through empathy or sympathy for the poster (e.g., “I know that must be really hard on you”). Words of Kindness are almost always combined with other response codes such as one of the Recommendation, Judgment, or Personal Experience; only one response expresses Words of Kindness alone. Adolescents thus rarely use Words of Kindness as their sole expression of support, but as a supplement.
Adolescents’ Responses to Different Romantic Difficulties
Our second research question concerns whether adolescents provide different responses for personal accounts of romantic difficulties of varying nature. We conducted pairwise chi-square tests to examine whether the distribution of the nine response codes differ across the four categories of romantic difficulties (i.e., CONTROLLING BEHAVIOR, BREAK-UPS, CRUELTY, and TRUST ISSUES) expressed in the data set. The proportions of response codes differ between the ABUSE category and each of the other three categories of romantic difficulties (p < .001). The overall proportions of the response codes are also different for CONTROLLING BEHAVIOR and BREAK-UPS (p = .001). However, there is no statistically significant difference in the proportions of response codes between CONTROLLING BEHAVIOR and TRUST ISSUES, and between BREAK-UPS and TRUST ISSUES.
Table 3 shows the proportions of the nine response codes by the four personal account categories of romantic difficulties expressed by adolescents on the ATL platform, along with 95% confidence intervals. Adolescent responders are more likely to suggest Seeking Help for the CRUELTY personal account category than for CONTROLLING BEHAVIOR, BREAK-UPS, and TRUST ISSUES. However, responders are less likely to suggest peers Communicate for romantic difficulties related to CRUELTY than for those related to CONTROLLING BEHAVIOR and TRUST ISSUES. In addition, responders are less likely to provide Judgments for romantic difficulties related to ABUSE partners than for those related to BREAK-UPS, and TRUST ISSUES. There are no significant differences in adolescent responses across the four different categories of romantic difficulties for Break-Up/Move on, Words of Kindness, Personal Experience, and Probing Question response types.
Proportions of Codes by Type of Romantic Difficulties.
Note. Confidence intervals indicated in the brackets are calculated at .05 significance level. The values outside the brackets indicate proportions.
Response Codes by Poster’s Gender
Our third research question concerns whether adolescents’ responses differ by the poster’s gender. To balance the difference in the number of responses to female (247 responses) and male (46 responses) posters in our sample, we conducted additional coding for all of the responses to male posters in the original data set (240 responses). Among the 247 responses to female posters, 74 (30%) are to posts about CONTROLLING BEHAVIOR, 51 (21%) to BREAK-UPS, 66 (27%) to CRUELTY, and 56 (23%) to TRUST ISSUES. Among the 240 responses to male posters, 29 (12%) are to posts about CONTROLLING BEHAVIOR, 138 (58%) to BREAK-UPS, 8 (3%) to CRUELTY, and 65 (27%) to TRUST ISSUES. We conducted chi-square tests for each of the 9 response codes by gender.
Chi-square tests for each of the nine response codes by gender show that female posters are more likely to receive Recommendations overall, and specifically Recommendations to Seek Help and Break-Up/Move on, than male posters (p < .001). The differences in distribution for other six response codes by posters’ age and gender are not statistically significant at the .05 level (see Table 4).
Distribution and p Value of Codes by Age and Gender.
Discussion
Romantic relationships represent an important context for adolescents’ development (Carver et al., 2003; Collins et al., 2009; Furman & Hand, 2006), and previous studies of online teen forums indicate that romantic issues are among the most prevalent topics for which adolescents seek support online (Weinstein & Selman, 2014; Suzuki & Calzo, 2004). Yet, many adolescents also indicate low satisfaction with online peer advice (Gould et al., 2002). The current study therefore specifically examines the content of authentic advice exchanged online in response to queries about romantic relationship issues. More generally, it adds to the growing literature on adolescents’ use of virtual environments for peer support and help-seeking. The findings also provide an empirically based entry point for considering both the nature and the implications of adolescent peer support online: They enable us to assess the contexts in which online advice is likely to help and support youth who are grappling with romantic relationship issues, when it may fall short, and how peer advice differs under different circumstance.
The anonymous adolescent responders on the ATL platform provide a range of responses that include not only—most often—Judgment or Recommendation but also Words of Kindness, Personal Experience, and Probing Question. Within the category of Recommendation, adolescent commenters’ strong preference for suggesting peers Break-Up/Move on over Seek Help, Communicate, or Other Recommendation is a noteworthy finding with mixed implications. Ending a problematic romantic relationship can certainly be an adaptive solution for adolescents (Tashiro & Frazier, 2003) that minimizes the time “staying-in” unhealthy relationships and end relationships before problem behaviors escalate (Helm, Baker, Berlin, & Kimura, 2015).
However, break-ups involve a complex and ongoing process that can be difficult to execute and can result in significant psychological distress (Field et al., 2009; Langhinrichsen-Rohling et al., 2000). In addition, given that adolescents experiencing abuse often hesitate to seek help due to fear that others will exert pressure to end the relationship (Mahlstedt & Keeny, 1993), receiving an abundance of resolute advice to cut ties and break-up might further isolate adolescents in relationship difficulties. Finally, the recommendations to Break-Up/Move on also consistently lack concrete steps for how to terminate a relationship, and do not point to potential sources of help for managing the aftermath. Although this pattern has also been documented in adolescents’ peer advice offline (Ocampo et al., 2007), the focus on breaking-up, along with the dearth of information on how to do so, likely limits the utility of these online recommendations.
The dearth of online peer recommendations to Seek Help also echoes previous findings on adolescents’ strong reluctance to seek help for romantic relationship difficulties (Molidor & Tolman, 1998; Ocampo et al., 2007). The absence of recommendations to Seek Help for CONTROLLING BEHAVIOR or TRUST ISSUES, which are related to psychological distress (Jouriles et al., 2009) and physical violence (Follingstad, Bradley, Laughlin, & Burke, 1999; Molidor, 1995), might be particularly unfortunate if it results in the loss of prevention or intervention opportunities for adolescents. In addition, the relative scarcity of Words of Kindness that directly communicate empathy, sympathy, hope, or encouragement indicates that while anonymous online forums might be useful portals to access peer feedback, they might not be robust pathways to the emotional support and empathetic feedback that adolescents value and crave from peers (Jackson et al., 2000). Thus, the types of recommendations adolescents exchange on the platform appear somewhat limited with respect to their adaptive value and utility, and in ways that seemingly reflect offline patterns.
However, adolescents treat accounts that describe CRUELTY differently from the other three types of romantic relationship issues—and in a manner that we interpret as a clear strength of the online peer exchanges in question. The CRUELTY personal accounts receive more recommendations to Get Help and to Communicate, and fewer Judgments, than other difficulties. Thus, adolescents on the forum seem to recognize and indeed provide different responses for romantic relationship difficulties anchored in explosive episodes of hostility and aggression, which likely constitute heightened risk. The CRUELTY posts also receive the second highest number of responses overall, which is noteworthy given that adolescents in the offline context exhibit avoidant tendencies in responding to peers’ relationship difficulties that are more severe (Weisz et al., 2007). Adolescent help-giving behaviors for these issues may differ online because both help-seekers and help-givers are more anonymous and are not communicating face-to-face.
Finally, adolescents’ response patterns also differ depending on the gender of the poster, and in a manner that we interpret as less beneficial. Boys are significantly less likely than girls to receive action-oriented Recommendations, specifically to Seek Help and Break-Up/Move on, as well as Words of Kindness. Notably, male and female posters to the forum did not share stories of each type of romantic relationship issue with comparable frequencies. The majority of responses to males in the sample are in response to posts about BREAK-UPS and rarely about CRUELTY, while the responses to girls are more evenly offered for all four types of issues. However, the prevalence of Break-Up/Move on recommendations and Words of Kindness do not differ between CRUELTY and BREAK-UPS for the total sample; thus, the naturally occurring skew in the prevalence of these two response codes is unlikely simply a product of gender differences in post topics. The lower prevalence of recommendations to Break-Up/Move on might indirectly perpetuate the tendency of males to accept and tolerate difficulties in relationships (Simon et al., 2001). Such concrete Recommendations from peers may be essential for boys as they report lower relationship confidence about breaking-up or communicating with a partner (Giordano, Longmore, & Manning, 2006).
The different prevalence of Seek Help recommendations for male and female posters may stem, in part, from the different post topics, because adolescents offer more Seek Help recommendation for CRUELTY than BREAK-UPS. Yet, boys almost never received recommendations to Seek Help. The absence of peer advice to Seek Help for boys is particularly noteworthy because adolescent boys are already unlikely to turn to others for help (Rickwood et al., 2005; Tishby et al., 2001) and struggle with burdensome gender stereotypes that define boys as self-reliant, tough, and in control (Addis & Mahalik, 2003).
In sum, our analysis indicates that peer advice for romantic issues exchanged in an anonymous online context mirrors several documented patterns of peer advice in the offline world. This includes advocating break-ups as a primary method of managing relationship challenges, rarely suggesting outside or adult intervention, and differential treatment of boys and girls. However, the benefits of these patterns are debatable, and peer advice online could, therefore, be more informed, detailed, and balanced.
At the same time, peer helpers online seem particularly sensitive to issues of relational cruelty: whereas adolescents offline often avoid interacting with these issues (Weisz et al., 2007), adolescents on the forum post more comments and adopt different patterns of advice for these stories, offering more recommendations to seek outside help. Thus, our findings suggest that, while moderated forums do not replace various sources of offline support, especially that of adults or professionals, they appear to provide a more inviting venue for adolescents to seek and receive peer feedback on serious issues that are hard to approach offline.
Our findings also identify areas in which additional scaffolding could improve the quality and utility of online peer support. Moderated platforms can prompt the types of response that are lacking in peers’ comments, such as detailed strategies for actionable recommendations, and Words of Kindness that might enhance the positive impact of online peer forums. These platforms can additionally provide information that is useful but consistently absent from peers’ posts, such as resources on romantic difficulties or links to verified professional services.
Online forums already mitigate a major barrier to help-seeking—fear of embarrassment (McKenna & Bargh, 2000; Suler, 2004)—by providing anonymous contexts to access support; the inclusion by the platforms of additional resources would alleviate yet another main barrier to help-seeking, which is lack of knowledge about where to go to seek help (Christiana et al., 2000). Finally, informing adolescents about the responses they might receive from peers online can equip adolescents with a more realistic expectation of the support they might receive; in society today, online peer support should be presented not as a replacement but a supplement to other types of offline support.
Limitations
One limitation of our study is the lack of complete demographic information for responders. Second, where demographic information is available, it reveals a considerable gender imbalance in this voluntary participant–based sample. Third, because of the self-selection of users into http://AThinLine.org—a website designed for teens interested in sharing personal experiences and helping others—our findings may not represent the responses of adolescents on other types of digital platforms. Moreover, the way the data were generated does not allow us to consider how adolescents interpret the quality or feasibility of the advice. For example, how useful is a response that advises communication with the partner such as “talk it out,” which does not reveal what kind of talking will be needed to help resolve romantic relational difficulties (nor what type of listening is also entailed)? We do not currently know whether or how adolescents act on these recommendations.
Directions for Future Research
The findings and limitations of the current study point to directions for future research to further improve prevention and intervention efforts. Investigations similar to the current study with other online platforms can help build a more comprehensive understanding of the types of responses to romantic relationship difficulties, as well as to other kinds of adolescent social challenges. Another crucial direction is to explore adolescents’ interpretations of different types of peer advice and support to understand which are experienced as most valid and useful. For example, what is the nature of adolescents’ receptivity and adherence to peer advice that suggests terminating their relationships versus seeking help from others versus ignoring the issue? In addition, future investigation—particularly with more gender-balanced samples—is vital to further clarify how and why adolescents respond differently to male and female posters. Finally, further research on the comparative strengths and weaknesses of online verves offline peer support is critical as adolescents’ access to online communities continues to expand. Supporting youth in their relationships requires continued research that carefully examines all the evolving contexts in which they seek support in the digital age.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.