Characteristics of romantic relationships during adolescence : A review of Anglo-western literature

Romantic relationships are a central aspect of the life of adolescents and have an important influence on the socio-affective and sexual development. This paper presents a review of the scientific literature on romantic relationships during adolescence and aims to unravel the ongoing debate in the literature and to identify possible gaps. The review encompasses the analysis of 36 studies on romantic relationships in adolescence, all published in peer-reviewed English scientific journals. The analysis of the published data was guided by five central characteristics of romantic relationships of adolescents: i.e., involvement, partner selection, content, quality, and emotions and cognitions emerging from the relationship. It was found that although during the last decade a growing interest of scholars in the study of romantic relationships in adolescents emerged, providing a certain body of knowledge on the topic in general, there are still gaps in the understanding of adolescents’ romantic relationships in different socio-cultural contexts. Limitations of the reviewed studies in terms of agreement on definitions, inconsistencies in sample population, research design and methodological approach are discussed.


INTRODUCTION
It has been argued that romantic relationships are a central aspect in the life of adolescents and an important characteristic of the socio-affective development during adolescence (Furman et al., 2002).
A romantic relationship is a social constructed concept that is being defined distinctively in history.
Currently, the most commonly accepted characteristics of adolescents' romantic relationships are referring to a voluntary mutual relationship where both members of a dyad agree that they date.It was Carlson (1987) (as cited in Jackson, 1999) who defined romantic relationships between unmarried people as dating or dating relationships.This would probably apply to all romantic relationships in adolescence.Later, the meaning of dating was broadened and considered the social process through which adolescents' practice and experiment with hetero-sexual relationships (Feiring, 1996).In this review the terms romantic relationship and dating will be used interchangeably.
Many authors have recognized the necessity of studying and understanding the characteristics of romantic relationships during adolescence (Brown et al., 1999;Collins, 2003), which is reflected in the increasing attention on the topic from both behavioral and social scientists.The increasing interest in adolescents' romantic relationships stems from the recognition that they are not simple or trivial.Indeed, different studies have found that romantic relationships are the most supportive relationships for female and male adolescents, and that romantic partners are a major source of support for many adolescents' especially in the late stage of adolescence (Furman and Buhrmester, 1992).
Currently, most studies on the topic are being conducted by developmental psychologists, who focus mainly on the age relationships start, the ways relationships cycle through time, and how these influence development (Furman et al., 1999).Collins (2003) proposed that the study of romantic relationships in adolescence should be centered around five features that characterize all close relationships, being: (1) the involvement of the teenagers in romantic relationships; (2) the process of partner selection during adolescence; (3) the contents of the romantic relationship of adolescents; (4) the quality of these relationships, and (5) the cognitive and emotional processes that emerge from these relationships.Apart from these categories proposed by Collins (2003), it is also important to consider gender differences and contextual factors in romantic relationships due to their salient importance in the development of romantic relationships.Indeed, one of the relevant aspects in the study of romantic relationships is that they are shaped by societal and cultural forces (Miller and Benson, 1999;Brown et al., 1999;Collins, 2003;Bouchey and Furman, 2003;Raffaelli and Ontai, 2004).
Given previous, the aims of the review presented herein are: (i) to summarize systematically the results of published studies in the Anglo-western literature on romantic relationships during adolescence; (ii) to analyze these results and unravel the ongoing debate in the literature on this subject; and (iii) to identify possible gaps in the area.In order to organize the research findings, the present review uses the theoretical framework as proposed by Collins (2003), both to guide the search for studies and to classify the results of the reviewed studies.

Retrieval
The first strategy applied to identify studies on the topic was to search relevant keywords that emerged from preliminary reading, being: romantic relationships, adolescence, socio-sexual development, socio-affective development, dating, gender, and violence.These keywords were combined with categories from the theoretical framework described above (involvement, partner selection, contents, quality, emotions and cognitions) in order to find relevant studies.The scientific articles were retrieved from different electronic search engines and databases, i.e.ERIC, Scholar Google, PubMed, and PsycInfo.

Inclusion criteria
To be included in the review, studies had to: (a) be published in a peer-reviewed English scientific journal, and (b) examine empirically the romantic relationships during the stage of adolescence.It was decided to only include studies about romantic relationships during adolescence in Western cultures because the lack of publications in the emerging and less developed countries, and the difference in cultural background resulting in a different interpretation of definitions about romantic relationships which would hinder the comparisons of results.Overall, 50 studies were found of which 36 met the inclusion criteria.These studies comprised the final sample.For the systematic systematization, analysis and interpretation of the data in the consulted studies a scheme was designed for their extraction and storage, being: title, location, date of publication, authors, objectives, hypothesis/research question, methodology, findings, conclusions and limitations.

CASE STUDIES CHARACTERISTICS
In order to identify trends in the research about romantic relationships in adolescence the characteristics of the studies are first highlighted.Focus thereby was on: the topics addressed the country where the research was conducted, the used sources, the characteristics of the target population, the methodological approach, the tools used and the main limitations reported by the scholars.
Topics.Emotional and cognitive processes emerging from romantic relationships appear to be the most frequently addressed in the consulted case studies (n = 10) together with the quality of romantic relationships (n = 11), while contents of romantic relationships is the topic least frequently examined (n = 4).Involvement in romantic relationships (n = 6) and the process of partner selection (n = 5) received attention as well.Since the topic is being studied mainly by developmental psychologists, it is evident that personal processes are the main point of interest in the area.Nevertheless, within the quality of romantic relationships, the topic of violence in romantic relationships gets quite some attention by scholars.
Place.Most studies in the survey report researches in the United States of America (n = 25).The rest of the studies were found to be conducted in Israel (n = 4), Spain and Canada (n = 2), South Africa, New Zealand, Australia and Germany (n = 1).

Sources.
The 36 analyzed papers were retrieved from 15 different peer-reviewed journals.Most studies were published in the Journal of Adolescence (n = 8) and Child Development (n = 4).These journals belong to different social disciplines as psychology, sociology and educational sciences.
Target population.The majority of consulted researches worked with samples of adolescents belonging to the different stages of adolescence (n = 12) while some of them focused on a specific stage of adolescence; with late adolescence being most frequently studied (n = 9) and early stage of adolescence being the least frequently examined (n = 3).Some studies were based on a retrospective search in a sample of young adults (n = 1) or adult women (n = 2).For such researches to be meaningful it is necessary to include in the analysis the probable impact of participant related factors, such as false memories and expectancy effects.Only a minority of studies (n = 11) included large samples (> 300 respondents).Most of the analyzed papers, however, addressed small school-based samples from specific geographic regions.Furthermore, a large fraction of the examined quantitative researches used data derived from broader studies covering regional or national samples.

Methods and tools.
Analysis of the used methodology revealed that a quantitative approach (n = 25) is most widely applied, whereas a few studies applied a qualitative approach (n = 5), and only one study a mixed quantitative-qualitative approach.The remaining articles in the survey (n = 6) consisted of a literature review of a given adolescence category.The quantitative studies mainly used surveys and questionnaires for data collection.Those instruments were specifically designed for the study of romantic relationships in the stage of adolescence, thereby using standardized self-report scales that were adapted from instruments designed originally for adults.The qualitative studies used in-depth interviews and non-participative observations for data gathering.It is remarkable that focus groups and narratives were rarely used.The mixed-method study used a survey in conjunction with complementary qualitative techniques to extend the explanatory power of their research findings.
When the focus in the study of romantic relationships is on the developmental perspective of adolescence a cross-sectional design (n = 26) was used compared to the longitudinal design (n = 7) targeting a comparison between the different stages of adolescence.

Reported limitations.
The main concerns expressed by the authors of the analyzed papers are related to size and composition of the sampled populations, with most studies based on small samples and limited to the heterosexual middle class of adolescents.Another important issue of concern is the necessity to conduct longitudinal studies, in order to understand the developmental patterns, the links between different categories and consequences during development.Finally, some authors worried about the limitation that studies were based on self-report questionnaires, possibly implying that the results rather may be reflecting what adolescents think instead of what they really do or feel about the topic.In fact, Feiring (1996) mentioned that during adolescence, there is a high tendency to give social desirable answers avoiding the discussion of negative or highly personal feelings or thoughts.Some authors suggested that in the analysis of romantic relationships, observational studies are needed to better understand factors associated with relational dynamics.Given the social constructive nature of the topic, contextual factors are frequently mentioned as a limitation for the generalizability of study findings as they indeed play an important role in the comprehension and understanding of the phenomenon.

CASE STUDIES FINDINGS
For the systematization of the results of the conducted survey of research papers, the main findings are presented in concordance of the five categories proposed by Collins (2003).

Involvement in romantic relationships
According to Collins (2003), involvement in romantic relationships refers to the questions about whether or not adolescents date and the age at which dating begins; the frequency and consistency of dating, and the duration of relationships.These aspects of involvement are considered to be indicators during adolescence.In this field, researchers focused mainly on three lines of analysis: frequency and duration of relationships, motives for romantic involvement and individual factors related to romantic involvement.In western cultures frequency and duration of romantic relationships are considered to be a salient feature during adolescence (Raffaelli, 2005;Shulman and Kipnis, 2001), in fact, studies developed with samples from different countries show that dating or having a romantic relationship during adolescence appears to be the norm (The Henry J. Kaiser Family Foundation/YM Magazine, 1998;Scharf and Mayseless, 2001), with a frequency that increases with age (Feiring, 1996;Shulman and Scharf, 2000;Carver et al., 2003).
The duration or stability of romantic relationships during adolescence is also extensively studied.Romantic relationships in adolescence can range from 1 to 60 months showing that even though sometimes they are neglected for being considered trivial and brief, they can be longer than expected (Feiring, 1996;Carver et al., 2003;Scharf and Mayseless, 2001;Karney et al., 2007).Feiring (1996) furthermore emphasized that although the length of dating relationships can be relatively brief, they can be intense in terms of frequency and contact.With respect to motives for romantic involvement, contrary to the stereotype that motives for romantic involvement for boys are related with an interest in sexual behavior, an American study discovered that boys engage in dating behaviors for a variety of reasons with relational aspects as the most important, suggesting an interest in their dating partner as a person and not simply as a sexual partner (Smiler, 2008).Other papers focused on identifying individual factors that are related to the onset of a romantic relationship.Individual qualities (sociability and physical appearance) and peer experiences (high quality friendships) were found to be predictors of the onset of romantic relationships (Zimmer-Gembeck et al., 2004).

Partner selection
Even though characteristics of romantic partners are considered to be an important factor in the development of romantic relationships, Collins (2003) highlighted that little is known about adolescents' choice of romantic partners or the consequences of romantic partner characteristics on psycho-social functioning.Studies with focus on partner selection have in general the selection of a sexual partner as main topic, in which often the main research goal is related with the spread of sexually transmitted diseases and HIV instead of the relationship itself (Andrinopoulos et al., 2006).Only few studies consider the selection of a romantic partner as a subject of research.However, despite the direct link between a romantic partner and a sexual partner, characteristics of one of the categories are not always possible to be extrapolated to the other, especially in cultures where a double standard is present; it is accepting that men are free to have premarital sex -even with multiple partners-while virginity and abstinence is supposed to be the rule for women (Sierra et al., 2007).
Based on the survey of the research papers with main focus on romantic partner selection, it is noticed that the main driver in the process of partner selection appears to be attraction depended on similarity.Homophily -the tendency for similar individuals to associate-has already been studied by social scientists in other types of relationships, mainly in friends' networks (Kao and Vaquera, 2006) and in adult couples (Urrea et al., 2004).A general tendency to be attracted to similar others increases the odds of choosing partners who share commonalities (Urrea et al., 2004;Simon et al., 2008).The homophily tendency is likely based on dimensions of status, including socio-demographic factors (age, ethnicity), as well as ascribed characteristics (social standing, physical attractiveness) (Simon et al., 2008).Age related homophily, however, is not considered trivial (Carver et al., 2003).Younger boys have been found to be paired with older girls, and older boys on average have been found to be paired with younger partners.Girls of all ages reported having romantic partners who were older on average.Besides the differences in age of partners, young girls dating with older individuals have been found to be more likely to engage in sexual intercourse in a relationship and this has been linked with risky sexual behaviors and with non-voluntary and forced sex (Manlove et al., 2005).
Nevertheless, contrary to the homophilic tendency, there is evidence that in some specific cultural groups, female adolescents tend to look for males who are from outside or from different groups.The former suggests that probably in some groups partner selection also implies possibilities for a change in social or economic status (Urrea et al., 2004).Another factor found to be important in partner selection is the presence of romantic love.In western societies, romantic love is related to an idealized conception of an asymmetrical relationship within a couple, where the role of a man is protection of his woman and offspring's, while the role of a woman is caring for the children and her husband (Urrea et al., 2004).
It is important to address two key issues in the process of partner selection: one is that features considered important may undergo developmental changes that have not been deeply explained by scholars.Some features (as e.g.physical attractiveness) may remain, while others (such as e.g.intimacy and conflict solving skills) may gain more significance (Urrea et al., 2004), and this mainly under the influence of the surrounding (i.e.parent and peers, but also social scripts).Gay adolescents for example may be particularly vulnerable to rejection from peers and parents because of the specific choice of romantic partner (Downey et al., 1999;Ford et al., 2003).

Content of romantic relationships
According to Collins (2003) content of romantic relationships refers to shared activities of romantic partners, what adolescent partners do together, how they spend their time, the diversity of their shared activities, and the activities and situations they avoid to do when they are together.It is, however, not easy to really know what adolescents do together within a romantic relationship.The best methodology would be to use observational techniques, but there is a lack of this kind of studies.Most of the studies are based on self-report and share the limitation of the possibility that adolescents answer questions rather according to what they think is supposed to happen or to what they believe to be the normative gender stereotyped behavior, instead of reporting what they really do based on an accurate assessment of their own behavior (Feiring, 1996).
Researches on the contents of romantic relationships have mainly focused on two main research lines: developmental changes (i.e.changes of patterns of behaviors) and types of relationships.With regard to developmental changes, romantic relationships are recognized to be guided by developmental principles, thus are expected to vary across ages (The Henry J. Kaiser Family Foundation/YM Magazine, 1998;Shulman and Seiffge-Krenke;2001). Romantic relationships during adolescence seem to occur in terms of phases that follow a specific pattern.Four stages have been identified: initiation, affiliative, intimate and committed stage.Such division by phases shows a clear evolution from short-lived attraction via casual interactions to deeper and more intimate and committed relationships (Shulman and Seiffge-Krenke, 2001).However, the models proposed do not establish any link between age and phase of romantic relationship.Shulman and Seifgge-Krenke (2001) indicated that romantic relationships have different meanings depending on the phase of adolescence and that adolescent romance moves from short-lived attraction via casual interactions with the other gender to deeper, more intimate, and committed relations.
Other studies revealed that in the development of romantic relationships romantic and social events occur before sexual events.Being together with one´s partner in a group and holding hands were reported to happen first, followed by private and social identification as a couple.Sex-related items: touching under or without clothes, talking about birth control and sexual transmitted diseases (STD's), touching genitals and intercourse were considered to be the last events in the patterns (O'Sullivan et al., 2007).Bauman and Berman (2005) stated at the other hand that in the United States three types of relationships can be discerned: messing, girlfriend-boyfriend and hubby-wifey.Differences between relationships are based on factors such as the degree of love and care, the extent to which sexual behavior implies a future commitment, the degree the relationship is public or a secret and the expectations of monogamy.
One important line of research in the study of contents of romantic relationships in adolescence is focusing on comparisons with other close relationships (e.g.Scharf and Mayseless, 2001;Furman and Shomaker, 2008).Studies in this line showed that romantic relationships share some common features with other relationships, but also have distinct features and patterns of interactions.Giordano (2003) for example made a comparison between romantic relationships and friendships: friendship is based on similarity and shared perspectives, heterosexual relations by their nature encompass differences; and romantic relationships are highly emotional, close friendships are generally considered a more settled and comfortable social arena.Whereas romantic relationships show great possibilities for asymmetries, friendships are characterized by symmetrical reciprocity.Furthermore, issues of exclusivity and commitment appear fundamental to understand romantic compared to friendship relations.

Quality
The quality of a romantic relationship refers to the degree to which the relationship provides generally beneficent/beneficial experiences.Collins (2003) stated that high-quality romantic relationships are those in which the partners manifest intimacy, affection, and nurturance; whereas partners in low quality relationships manifest irritation, antagonism, and notably high levels of conflict and/or controlling behavior.In fact, romantic relationships in adolescence have the potential to contribute to positive personal and interpersonal growth, however, they are also potentially a space for risk behaviors and dysfunctional relationships, which can affect a healthy emotional growth of the adolescent (Giordano, 2003).
Researches on the quality of romantic relationships in adolescence mainly focused on two areas: (a) the identification of the developmental role of romantic relationships for adolescents with eventual positive and negative influences, and (b) the identification of the features that categorize romantic relationships as healthy or non-healthy.Several scholars emphasize that romantic relationships may play an important role in the psycho-social and affective development during adolescence (Furman and Shafer, 2003); underlining the benefits that dating could contribute (Bouchey and Furman, 2003).In fact, dating may serve diverse purposes, some of which overlap with friendship, including reaction, seeking autonomy, seeking status, sexual experimentation, and social skills practice and experimentation (Feiring, 1996).As stated in other researches it is difficult to discern the difference between healthy and non-healthy relationships because of the individual and interpersonal processes they involve.Some romantic relationships could include a certain degree of struggle and unhappiness, while many dysfunctional romantic relationships may include some positive elements.Larson et al. (1999) suggested that both positive and negative emotions emerging from romantic relationships can be an arena for psychological maturation, depending on features as state of consciousness.According to Sorensen (2007) healthy romantic relationships are characterized by open communication, high levels of trust, and partners who are close in age.This author states that healthy relationships help adolescents to refine their sense of identity, assists in the development of personal skills, and provides emotional support.
In an attempt to identify indicators of interpersonal dysfunction of adolescents romantic relationships, Florsheim and Moore (2008) conducted a group comparison observational study with adolescents that were previously classified as high and low risk (in high-risk group were included the adolescents who presented disorders of substance abuse and substance dependence).Results of this study suggested that higher levels of hostility may be a useful marker of dysfunctionality and predictor of other extra-relational risk behaviors (e.g.substance abuse or other types of psychopathology).The level of interdependence or enmeshment seemed not to be a good indicator.
Violence is considered as a main indicator of the unhealthy character of a relationship, and it has become an important topic for researchers who recognize the complexity of the phenomenon as well as the magnitude of impact it can have in the lifetime of individuals (Centers for Disease Control and Prevention, 2009).According to the Office on Women's health of the US Department of Health and Human Services (2009) dating violence is when one person purposely causes physical or psychological harm to another person they are dating, including sexual assault, physical abuse and psychological/emotional abuse.Studies addressing dating violence focus most frequently on the prevalence of the phenomenon, the risk factors, gender differences and contextual factors that are related with it.Recent studies suggest that in the stage of adolescence dating violence is common.Coincidences in prevalence rates of violence in different populations confirm that the phenomenon of dating violence is not local or specific for a population, instead it appears to be a universal problem (Halpern et al., 2001;Swart et al., 2002;Molidor and Tolman, 1998;Muñoz-Rivas et al., 2007) with a prevalence that increases as adolescents' age, but starts declining at around the age of 18 (Carver et al., 2003).
Dating violence has been found to be also common between adolescents involved in same-sex intimate relationship and it seems that the prevalence in these groups is similar to that of the heterosexuals (Freedner et al., 2002;Halpern et al., 2004).Furthermore, according researches both male and female adolescents can be either victims and/or perpetrators (Swart et al., 2002;Molidor and Tolman, 1998;Muñoz-Rivas et al., 2007).Due to the complexity of the phenomenon, many scholars focused their studies on identifying risk factors for dating violence (Young and Furman, 2007).Vézina and Hébert (2007) published a literature review on violent victimization risk factors for adolescent girls and young adult women in romantic relationships, and they identified four categories of risk factors: socio-demographic factors, individual factors (personal and interpersonal), environmental factors (family, community, and peers), and contextual factors (linked to the romantic relationship).
Gender differences in dating violence have also been the subject of several investigations.Despite similarities found in the rates of dating violence for females and males -suggesting that violence is similar across genders (Smita Varia, 2006)-, important differences have been found with regard to the kind of violence males and females are confronted with.Significant differences by gender in the type, severity, and seriousness of the consequences of violence were identified (Halpern et al., 2001;Swart et al., 2002;Molidor and Tolman, 1998;Bentley et al., 2007;Muñoz-Rivas et al., 2007), as well as gender based differences in the meaning attributed to a partner's act of violence (Swart et al., 2002) and the motives for the aggression (Muñoz-Rivas et al., 2007).Contextual factors seem to play a key role in the phenomenon of violence due to the fact that some cultural beliefs may promote or justify violence between partners (Valls et al., 2008).In this regard it is suggested that dating violence should be studied from an ideographic or situational perspective.
The analysis of the consequences of violence led to the identification of both short term (school problems, drugs or alcohol use, depression among others) and long term consequences (patterns of abuse might be carried over to future relationships), suggesting that dating violence could affect the person for lifetime (Smita Varia, 2006;Vézina and Hébert, 2007;Centers for Disease Control and Prevention, 2009).

Cognitive and emotional processes
According to Collins (2003) for the characterization of romantic relationships it is necessary to consider the distinctive emotional responses, perceptions, expectancies, schemata, and attributions that emerge within a relationship regarding oneself, the other person, and the relationship itself.Emotions and cognitions are closely intertwined and play a major role in determining their functional significance.When focusing on emotions and cognitions in romantic relationships during adolescence most current researches are theoretically based on concepts of the Attachment Theory and the conceptualization of love as an attachment process (Hazan and Shaver, 1987).In fact, three main categories of Attachment Theory are classically used to explain the development of cognitions and emotions regarding romantic relationships, these are: working models, behavioral systems and attachment styles.
The representations or expectations about the self, significant others, and the relationship between the two, have been conceptualized by attachment researchers as working models (Pietromonaco and Barrett, 2000).According to attachment theorists, working models as cognitive representations may constitute a positive or negative model of the other (based on the availability and responsiveness of significant others) and a positive or negative model of the self (based on the view of the self as competent).Attachment theorists also state that working models, once formed, continue to function as relative durable schemata for later relationships during lifetime.Based on Bowlby's conceptualization, Hazan and Shaver (1987) conceptualized romantic love as an attachment process.They suggested that individual differences in how romantic love is experienced may be due to differences in past attachment history and working models developed during earlier years.
Over the years the term working models has been replaced by terms as schemas, scripts, prototypes, and states of minds.Currently, views is a widely used term as proposed by Furman and colleagues (Furman and Wehner, 1994;Furman and Simon, 1999), referring to the mental representations that involve feelings and beliefs.In concordance with the conception of working models, in theory views stem from previous experiences in close relationships with parents and caregivers.Researchers, however, suggest that in the stage of adolescence, romantic views are expected to be developed with the contribution of relationships with friends, ongoing experiences with romantic partners, models of other romantic relationships (e.g.parent's marriage) and cultural expectations (Simon et al., 1998;Connolly and Goldberg, 1999;Furman, 2002;Bouchey and Furman, 2003).
Based on the conceptualization of romantic love as attachment process and the recognition of the existence of three behavioral systems -attachment, care giving and sex-, in the search of the understanding of romantic relationships during adolescence a fourth behavioral system affiliation was included (Furman and Wehner, 1997;Feiring, 1999).Elucidation of the existence of different behavioral systems can be found in the explanation of social networks and the organization of relational views.During lifetime, individuals have social networks in which different people serve as a main figure for the different behavioral systems: there are attachment figures; affiliative figures and individuals to whom care is provided.These individuals are hierarchically organized according to their importance to the adolescent.The idea of a hierarchical model was also used to explain the existence of different levels of relational views (Furman and Simon, 1999).According to these authors, views are organized in different layers, thus individuals have views of relationships in general, views of types of relationships, and views of particular relationships.According to this concept it is likely that other types of relationships can influence romantic views, create the opportunity for discontinuity between parent-child and romantic relationships.Furman et al. (2002) further suggest that the different layers of the hierarchy to be interdependent and that they are not necessarily concordant.
Based on the comprehension of attachment styles, Furman et al. (1999) proposed that romantic views can also vary from person to person and that those individual differences in relational views can be captured by the same categorical system used by attachment researchers: secure, anxious-avoidant (dismissing), anxious-ambivalent (preoccupied), and disorganized/unresolved (Furman et al., 2002).Nevertheless, categories of romantic styles are supposed to capture differences not only in mental representations of attachment but also the mental representations of the behavioral systems operating in a given relationship (Simon et al., 1998).In concordance with the measurement of attachment styles, romantic styles can be measured by interviews and self-report techniques.These measurements include also the assessment of intimacy and closeness with respect to care giving, affiliation, sexuality and attachment systems.The distribution of romantic styles among the adolescents seems to be similar to the distribution of the attachment styles in adults (Seiffge-Krenke, 2006).An important line of research regarding romantic styles focuses on the identification of links between romantic styles and other variables such as: behaviors (Furman and Simon, 2006), psychological health (Dávila et al., 2004;Seiffge-Krenke, 2006), sexual satisfaction (Brassard et al., 2007), among others.
It is important to mention that among the five features analyzed, in the comprehension of emotions and cognitions emerging from romantic relationships, it is relevant to consider the developmental perspective, gender differences and contextual factors, which all have an important impact in the conformation of emotions and cognitions.In line with the developmental perspective of the study of romantic relationships in adolescence, romantic views are believed to change across age.In order to identify developmental changes in the perceptions of the social networks, Furman and Buhrmester (1992) developed a cross sectional study with adolescents from early, middle and late adolescence, as well as with young adults.The authors found that the main propositions derived from the major theories of the developmental courses of personal relationships could be supported, these are: (a) mothers and fathers were seen as the most frequent providers of support in the early adolescence, and (b) same-sex friends were perceived to be as supportive as parents in the middle adolescence, but became the most frequent providers of support in late adolescence.
Regarding perceptions of romantic partners, the analyzed publications show that they move up in rank with age until young adulthood, where they, along with friends and mothers, receive the highest ratings for support.In the same line, Shulman and Scharf (2000) found that some perceptions of romantic relationships remain similar across adolescence, while others differ across ages.These authors identified that experience of romance as a care giving relationship as well as the value placed on intimacy and attachment was predominant across the various stages of adolescence, while younger adolescents emphasized aspects of friendships or companionship.This tendency was lower during late adolescence.A decrease of the perception of excitement related to romantic relationships was also detected in this study.Regarding gender differences, romantic views are reported to differ according to gender in different ways such as: descriptions of romantic relationships (Feiring, 1996); expressing interest in the other sex (Feiring, 1996); working models styles distribution (Gurza Ruiz, 2004;Furman and Simon, 2006); and perception of social networking (Feiring, 1999;Furman and Buhrmester, 1992).
Most authors emphasize the impact contextual factors have on the construction of romantic views through different mechanisms in conscious and unconscious ways.Bouchey and Furman (2003) stated that romantic relationships occur within a social context and that consequently adolescents' romantic experiences and views are linked to their close relationships with parents, friends, and romantic partners.Besides these close relationships, there are also media portrayals of romance and norms about dating, sexuality, gender roles and marriage that may influence romantic relationships and expectations.These authors emphasize that considerable variability exists across cultures.In the same line, Giordano et al. (2006) underlined the influence cultural expectations can have on emotionmanagement and emotional-expression of romantic views.Since different patterns of socialization of boys and girls may have an important influence on romantic views, several studies focused their attention on those processes (Rafaelli and Ontai, 2004;Feiring, 1996).Women are socialized, to be more relationship and care giving oriented and expressive of love.Based thereon, Raffaelli and Ontai (2004) suggested that research should directly assess relevant cultural (e.g.values and beliefs) and demographic variables, and examine how both sets of variables operate on family and gender socialization.
Finally it is important to address that even though attachment theory is the dominant perspective in the explanation of romantic relationships, when working with adolescents some authors criticize it as a main framework.Giordano (2003) argued that attachment models are not entirely comprehensive as a framework for understanding adolescent relationships for three main reasons.First, he states that the notion of carry-over across relationships fails to adequately highlight the unique developmental roles, subjective meanings, and relationship dynamics connected to each form of relationship.Second, he argues that the notion that attachment is a priori a developmental good constitutes a limited conceptual bridge to the types of problem outcomes in which sociologist are interested in.Third, this author highlights that attachment approaches tend to be individualistic, which constitutes a limitation for explanations of broader social influences on a relationships form or content.

CONCLUSIONS
The literature review presented herein aimed to provide a state-of-the-art about scientific research on romantic relationships in the stage of adolescence.The analysis of the selected scientific material allowed reaching two main targets.First, the main findings from the studies enabled to build a body of knowledge about the main features of romantic relationships during the stage of adolescence.Second, the analysis of methodological aspects of the reviewed studies resulted in the identification of existing gaps and limitations, and recommendations for future research.
Even though the current literature review has allowed building a broad picture of the features of romantic relationships in adolescence, it is important to consider the effect of limitations in used methods on obtained results, as summarized in the following.

Lack of standardized definitions
One of the main limitations of research in this area is related to the lack of standardized definitions about romantic relationships.The latter is the consequence of the difficulty to provide a unique and standardized definition of romantic relationships due to cultural differences in conceptions, but also due to the differences within individuals from the same cultural group.By lack of universal definitions, studies are based on very different questions hindering the transversal analysis of publications.Karney et al. (2007) also reported the lack of precise terms as a main obstacle for the study of romantic relationships.They stated that researchers tend to leave the task of definition to the study participants, by giving them the decision whether and when they are in romantic relationships.In her literature review, Jackson (1999) also expressed concern about the failure of defining what is meant by dating and proposed that confusion about dating could be avoided if researchers would make explicit their operational definition of dating for any given study.Unfortunately, even today most researchers do not pay sufficient attention to this issue.

Sampling considerations
Most of the studies are based on small samples of white, middle class, heterosexual adolescents.There is an absolute lack of research about same sex oriented adolescent couples.For future research, it will be necessary to broaden the studies to representative samples from different ethnicities, cultures, countries, groups, and sexual minority youth groups in order to capture the diversity of expressions and to give a better comprehension about how cultural and demographic variables may influence relationship processes and their expressions.This implies that current results have restricted transferability to other populations, because contextual factors play an important role in the comprehension and understanding of the phenomenon.More studies with bigger and more diverse samples are needed to allow the findings to be generalizable to groups that share common features.

Nature of research design
Even though the study of romantic relationships during adolescence is being conceptualized based on a developmental perspective, there is a lack of longitudinal studies.Most results are based on crosssectional studies and comparisons among stages.Karney et al. (2007) mentioned the weaknesses of cross-sectional designs for drawing inferences about causal influences.Future research should consider longitudinal studies that may give a better explanation about the developmental course of adolescents' relationships as well as of the developmental patterns of cognitions and emotions.Finally, the lack of observational studies implies that most findings are based on self-report studies implying the limitation that adolescents may inform what they consider socially desirable.During adolescence, there is a tendency to present socially desirable information and inhibit the discussion of negative or highly personal feelings or thoughts, which implies that results may reflect what adolescents think it should be instead of what they are really doing, living, believing or feeling.On the other hand, observational studies imply a greater effort in terms of time and resources.An additional important limitation is that not all the analyzed researches include the findings of both partners, thus results are mainly based on the report of one partner, including just one view.

Lack of qualitative studies
There is clear a lack of qualitative studies that could increase the explanatory power of results.Due to the features of the subject, quantitative studies with different samples are important, but qualitative studies are needed to get a deeper insight and explanation of feelings, emotions, personal processes and conceptions.A mixed-method approach seems to be necessary in order to acquire better understanding from the combination of results, as well as to harness improved knowledge about the differences between groups and the comprehension of the phenomenon in particular contexts.
In conclusion, although there has been a growing interest in the study of romantic relationships in adolescents in the last decade, which resulted in a certain body of knowledge on the topic in general, there are still gaps in the understanding of these relationships, particular in socio-cultural groups and contexts different from the one examined.Findings in the present literature review might be helpful in defining an approach to analyze the importance and characteristics of romantic relationships during adolescence in other socio-cultural environments, such as for example the Latin American civilization.