Note from the Field: Reflecting on Romance Novel Research: Past, Present and Future
by A. Dana Ménard
Abstract: In 2011, I co-authored an article in the journal Sexuality & Culture describing a study that I had done on sexual scripts in romance novels entitled, “‘Whatever the approach, Tab B still fits into Slot A’: Twenty years of sex scripts in romance novels” (Ménard & Cabrera, 2011). Shortly after the article appeared, a discussion about the paper took place on “Teach Me Tonight,” an academic blog devoted to the study of popular romance novels. The goal of this article is to further that discussion by exploring previous research on romance novels in the social sciences, explaining the rationale behind the methodology of my study, critically evaluating that study, and making suggestions for future work by romance novel researchers.
About the Author: A. Dana Ménard is a senior PhD candidate in clinical psychology at the University of Ottawa and is currently completing a one-year residency at the London Health Sciences Centre. She would like to thank Shannon Lawless, M.A., Vicki So, B.J., Chelsea Honeyman, Ph.D., and Christine Cabrera, M.A., for their assistance in strengthening this paper.
In 2011, I co-authored an article in the journal Sexuality & Culture describing a study on sexual scripts in romance novels. The paper was entitled, “‘Whatever the approach, Tab B still fits into Slot A’: Twenty years of sex scripts in romance novels” (Ménard & Cabrera, 2011). Shortly after the article appeared, a discussion about the paper took place on “Teach Me Tonight,” an academic blog devoted to research on romance novels. Romance novel researchers and fans offered their comments, questions and critiques on the study, and the editor of the Journal of Popular Romance Studies invited me to write a piece reflecting on [End Page 1] my original article and on the exchange—in particular, on the methodological and disciplinary issues that seemed to be at stake.
The purpose of this paper, then, is to continue and elaborate on that discussion, with specific references to the original study and to the dialogue that was generated from the “Teach Me Tonight” post. My goals are to provide additional details about how that study was developed, so as to clarify and contextualize the findings from the research, and to make some suggestions, based on my own limited experience, for how romance novel researchers might proceed going forward. To that end, this paper will include an evaluation of previous research on romance novels, an explanation for the rationale behind the methodology of the sexual script study, a critical evaluation of the strengths and limitations of that study, and some suggestions for what might be done by future romance novel researchers.
I am currently a senior PhD student in clinical psychology. I came up with the idea for the original study in my second year of the program, conducted the bulk of the research during my third and fourth years, and am writing now from the vantage point of my sixth and final year. My primary area of research since 2005 has been human sexual functioning, with a specific focus on optimal sexual experiences (e.g., Kleinplatz & Ménard, 2007; Kleinplatz et al., 2009a, b). My secondary area of interest is in depictions of sex and sexuality in the media. Prior to undertaking an investigation of sex and sexuality in romance novels, I had researched and co-authored a study on depictions of “great sex” in lifestyle magazines (e.g., Cosmopolitan, Men’s Health) (Ménard & Kleinplatz, 2008). I am also currently involved in a project investigating depictions of gender roles, sex and sexuality in “slasher” movies (Weaver, Ménard, Cabrera & Taylor, in press). Although I have conducted and published research on romance novels, I do not identify as a romance novel researcher, per se. My research interests in this area are focused on the study of sex and sexuality within the context of psychology, and I chose to investigate romance novels as an example of a widespread, popular media.
Goals of the original study
My goal in undertaking this research was to add to the body of knowledge concerning the content of media depictions of sexuality, as well to gain additional experience designing and conducting research, with an eye to publishing the findings in a peer-reviewed psychology journal. The study was not undertaken in order to receive course credit, but was, truly, a labour of love for both myself and my original co-author. I have long felt that researchers in psychology tend to focus on unusual or outlying human experiences—especially on those that are negative or unpleasant. I am personally more interested in the idea of studying everyday life experiences in order to develop a deeper understanding and appreciation of them and of what it means to be human. Reading romance novels is an experience that many people in the general public share; therefore, I [End Page 2] thought that this study might be relevant and interesting to people outside the world of academic psychology. On a more personal level, I was raised in a home where romance novels sat on bedside tables and bookshelves; I read many of them myself, including many by Nora Roberts. My mother, grandmother, and aunt all traded romance novels back and forth among themselves and their friends, a practice that continues to this day.
Previous researchers have tended to focus on the degree to which individuals’ beliefs about sex and sexuality are consistent with dominant sexual scripts; however, personal attitudes and beliefs in this area are not created in a vacuum, but, rather, are influenced by the wider cultural context. Research in psychology on representations of sex and sexuality in the media has tended to focus on the impact of consuming such material (e.g., Brown et al., 2006; Kim & Ward, 2004; Pardun, l’Engle & Brown, 2005) rather than the content of the material itself. This is a problematic omission and has created a situation where we know that people are impacted by media messages but we do not know much about the messages themselves. It was hoped that the results of this research would extend previous findings on sexual scripts as well as add to the literature on media representations of sex and sexuality.
The goals of the original study (i.e., Ménard & Cabrera, 2011), which will be referred to in this paper as the “sexual script study,” were, first, to gain an understanding of how sex and sexuality are portrayed in single-title, contemporary romance novels and, second, to determine whether these portrayals had changed over the last 20 years (i.e., from 1989 to 2009). It was hypothesized that most depictions of sex and sexuality in romance novels would adhere to traditional Western sexual scripts, and that adherence rates would not change significantly over time. Gagnon and Simon, the developers of Sexual Script Theory (SST), state that the purpose of sexual scripts is to identity the “who, what, when, where and how” elements of a sexual encounter (Gagnon, 1977; Gagnon & Simon, 1973; Simon & Gagnon, 1986, 1987). These scripts allow people to identify appropriate sexual partners, times, places, behaviours, and sequences, based on cultural norms (Wiederman, 2005). For example, men in Western culture are thought to focus almost exclusively on sexual pleasure and experimentation while women are thought to demonstrate a greater focus on the romantic context for sexual activity (Wiederman, 2005).
It was hypothesized that the male and female protagonists in the study sample would be young, attractive, Caucasian, able-bodied, heterosexual and single. Couples would not demonstrate obvious discrepancies on any of these variables (i.e., one partner bisexual and the other heterosexual, or one character able-bodied and the other disabled). It was expected that most sexual encounters would be initiated by the hero and would take place in a bedroom, at night. Sexual behaviours were expected to follow a particular sequence, i.e., kissing, touching, manual/oral sex and penile-vaginal intercourse to orgasm, within each scene as well as across the book (e.g., couples would kiss earlier in the book prior to engaging in oral sex). “Kinky” sexual behaviours (e.g., BDSM-inspired, threesomes, and anal sex) were unlikely to occur (where the word “kinky” is used in reference to dominant sexual scripts rather than to the personal sexual values of the co-authors). Finally, it was predicted that books published from 1989 to 1999 would not differ significantly on any of these variables from books published from 2000 to 2009. [End Page 3]
Description of the original study
To evaluate these hypotheses, a content analysis was done on sex scenes and characters from the 20 most recent winners of the RITA award for best single-title contemporary romance novel. A coding form and a coding manual were created in collaboration by the two co-authors based on readings in SST. Pilot coding was done using a book from the research sample as well as several books outside the sample. Both researchers read each of the 20 books included in the sample to identify and code relevant scenes. Researchers met frequently throughout the process in order to ensure comprehensiveness and consistency in coding, with the goal being complete agreement on coding. This resulted in a final sample of 46 scenes and 44 characters, where the scenes and characters represented the “units of analysis” for statistical purposes. Variables were coded categorically in most cases. The results from the first major hypothesis were presented using descriptive statistics, i.e., the percentage of characters or scenes conforming to pre-study predictions. T-tests, with an alpha level set at .05, were used to determine whether there were differences between the sexual behaviours of male and female characters. The second hypothesis, whether there were differences between novels published from 1989 to 1999 and those published from 2000 to 2009, was also evaluated using T-tests, again with an alpha level set at .05.
The results showed that the original hypotheses for the study were supported with respect to characterization of the male and female protagonists, characterization and context of the romantic relationships, and order and nature of sexual behaviours. Specifically, romance novel characters in the novels studied—a sample drawn from the 20 most recent winners of the RITA award for best single-title contemporary romance novel, as explained above—were consistently attractive, Caucasian, heterosexual, single and young; 86% of heroines and 77% of heroes fit this description perfectly. The majority of romance novel relationships did not involve a significant discrepancy in terms of these variables (77%); that is, within a relationship, both characters were attractive, both were Caucasian, both were heterosexual, etc. Most of the time, sexual behaviours occurred in the “correct” sexually-scripted order across the book (77% of the novels) as well as within any given sex scene (90%). In addition, sex scenes never included “kinky” elements, such as anal stimulation or BDSM-influenced behaviours. A surprising finding was that, despite the prevailing stereotypes around romance novels in the popular media, very few of the scenes (17%) included “romantic” scene-setting elements, such as flowers, candles or lingerie. There was more variation in terms of location and time: 65% of scenes took place in an “appropriate” location (i.e., a private place, such as a bedroom) and slightly more than 72% took place in the evening or at night. More than half of the sexual encounters were initiated by the male character (54%), followed by female initiation (33%) and “initiated simultaneously” (13%).
In terms of the second hypothesis, there were no significant differences between books published between 1989 and 1999 and those published from 2000 to 2009 in terms of protagonist characterization, relationship characterization, order of sexual behaviours across the book, order of sexual behaviours within sex scenes or location and timing of sexual activities. The only finding that reached statistical significance was an increased [End Page 4] usage of contraception by characters in the second time block (from 18% of the scenes in the first time block to 58% in the second).
Evaluating Past Romance Research
To provide a context for the design of the sexual script study, the methodologies employed by other romance novel researchers will be described and critiqued. The focus will be primarily on romance novel research within psychology and the social sciences, as this has been the sole focus of my academic training. I believe it would be intellectually dishonest of me to evaluate research on romance novels from other disciplines (e.g., English), as I am not familiar with the standards or expectations within other academic traditions. I could not evaluate these writings in an informed manner and accord them the respect that they deserve. In addition, given the different norms between the two fields, it seems meaningless to compare and contrast literary criticism to social science studies. It may be noted that some studies are cited here that were not included in the publication of the original study. In some cases, the omission was made because the results from these papers were considered “outdated” by psychological standards (i.e., 30+ years old), and the specific findings may be less relevant in a world where sexual mores are rapidly evolving, even over the last five years (e.g., the advent of “sexting”). However, I have included these studies here because they were influential in the design of the sex script study as well as the selection of research questions and variables of interest. Likewise, since the development and publication of the sex script study, which began in the fall of 2008, a few more articles about romance novels have appeared in scientific journals (e.g., Cox & Fisher, 2009; Fisher & Cox, 2010). These papers have also been included in an attempt to provide the most comprehensive overview of different methodological approaches to romance novel research and in the hope of suggesting some direction for future romance novel researchers.
Within the social sciences, research on romance novels has tended to center on how gender roles, sex and sexuality are depicted in these books. More specifically, researchers examining gender roles have looked at romance characters’ professions, personal qualities, family roles, and physical appearance characteristics (Clawson, 2005; Ruggiero & Weston, 1978). Romance novel researchers concerned with portrayals of sex and sexuality in these books have studied contraception usage, sexual behaviours, and sexual communication and initiation (Diekman et al., 2000; Thurston, 1987). With certain exceptions (e.g., Cox & Fisher, 2009; Fisher & Cox, 2010), romance novel researchers tend not to adopt a specific theoretical framework to guide the design of the study or the evaluation of the results. The content analysis of films, television shows, music videos and books has a long history within research on sex and sexuality; however, it is less easy to explain and understand why other important phenomena (e.g., parenting, mental health, friendship) have not been studied within the context of romance novels. If media represents a mirror to predominant cultural beliefs, there are no limits to what might be interesting and useful to investigate through the lens of romance novels.
Studies on romance novels in the social sciences have varied significantly in terms of sample size and composition. Researchers have chosen to investigate as few as 24 novels [End Page 5] (Weston & Ruggiero, 1978) or as many as 120 (Clawson, 2005). Cox and Fisher (2009) included 15,019 romance novels in their study, but they focused exclusively on words appearing in the titles of these books. It should be noted that virtually all researchers have constructed their samples using entire romance novels rather than scenes or characters from these books. This may represent a limitation to earlier studies because the use of entire novels may obscure differences between various sex scenes in the novel. Romance novel authors have multiple opportunities to reinforce a particular message (e.g., “real sex” must involve intercourse, heroes must be hyper-masculine) or to introduce varied approaches to the same idea (e.g., “real sex” may involve a variety of behaviours, heroes can be androgynous), if they wish. Therefore, analysis at the level of the scene may be a more precise measurement of the messages contained in the novel. Within the field of content analysis, sample sizes are generally smaller than population-based research; a sample size of 30 units, where “units” are defined as novels, magazines or films, is sometimes recommended (Lacy & Riffe, 1996). It should be noted that a small sample size does not represent an inherent limitation to a study, as specific statistical analyses have been designed to correct for different sample sizes.
Romance novel researchers in the social sciences usually reported the range of publication dates for the novels (e.g., from 1960 to 1980) but rarely provided additional details, such as the number of books selected from each year. This is a concern because it is unknown whether the sample included an over-representation of books from a particular era (e.g., the more sexually-conservative 1980s). In addition, previous romance novel researchers have not studied changes over time with respect to their research questions. Again, this seems surprising as romance novels, like other forms of media, both shape and reflect social changes. Romance novel authors are raised with a particular set of cultural expectations around sex and sexuality, which are shaped by the media, family, friends, teachers, etc.; their personal beliefs about sex may then influence how they choose to write about this. In turn, their books may then have an impact on the beliefs and values of their readers. Authors may also be influenced by the guidelines that they receive from publishers and editors, as well as from the traditions or conventions of the particular sub-genre; these guidelines may evolve over time as the publishing houses react to and shape changes in social mores. Whether or not changes in romance novel parallel social changes over time is an important question and one that future researchers would be advised to consider when constructing their research samples.
Within the social science literature on romance novels, the criteria used to identify books for inclusion in the study sample varied considerably. Some authors focused on popular authors within a sub-genre (e.g., Ruggiero & Weston, 1978), while others focused on specific publishers or specific lines within a publishing house, where the publisher, in most cases, was Harlequin (e.g., Clawson, 2005; Cox & Fisher, 2009; Fisher & Cox, 2010). It is likely that selecting only books published from one specific house may limit the applicability of findings from that study to books from that publishing house alone. Similarly, the focus on a particular subset of authors may illuminate commonalities in those authors’ books but may limit applicability of findings to those specific authors. However, in some cases, limiting the focus of the sample may be a desirable strategy as certain publishing houses or authors may have a proportionately greater influence both on readers and on the stylistic choices of fellow authors. Diekman et al. (2000) randomly selected every third “modern romance novel” at three different bookstores, which is the strategy [End Page 6] that methodologists believe is most likely to lead to generalizeability of study findings. However, the random selection of romance novels from a bookstore or library may result in over- or under-representation of particular authors or sub-genres, depending on the taste of the store manager or librarian. So far, no researchers in the area of romance novels have used bestseller lists to select novels (e.g., the New York Times, Publisher’s Weekly), although this strategy has been used to study other kinds of books (e.g., self-help books; Zimmerman, Holm, Daniels & Haddock, 2002). Each of the approaches listed above are viable and informative as long as researchers are clear about the generalizeability and limits thereof of their findings.
As expected, the methods that researchers used to select romance novels and the inclusion of romance novel sub-genres were correlated. Some researchers chose to deliberately focus only on specific subgenres, such as Christian romances (Clawson, 2005), erotic/historical romances (Thurston, 1987) or gothic romances (Ruggiero & Weston, 1978). In their study of Harlequin romance titles, Fisher and Cox (2009, 2010) included books from many different sub-genres (e.g., historical, intrigue, blaze). In some cases, the composition of the study sample with respect to sub-genres is unclear (e.g., Diekman et al., 2000). The decision to focus specifically on one sub-genre or to mix multiple sub-genres in one study has important ramifications for the interpretation of results. The applicability of findings from studies focused on one sub-genre will be limited to that sub-genre alone, but results from those studies may result in a deeper, more comprehensive understanding of that area. If the study sample includes a variety of sub-genres, the results may represent a better picture of the state of romance novels as a whole but may lead to a distorted picture of the “average” romance novel (if such a novel could be said to exist). Given the diversity of books included under the more generic “romance novel” umbrella, it seems unlikely that the results of research on one sub-genre of romance novel could be generalized with any accuracy to another sub-genre. Romance novel researchers who include multiple sub-genres in their study sample would be advised to focus on a few sub-genres, and to analyze findings separately before pooling results.
Evaluating the Sexual Script Study
In April 2011, I stumbled upon the discussion on Teach Me Tonight of the study I co-authored. The tone of the discussion seemed more elevated and respectful than is typical of many Internet debates, and I felt welcome to participate and to try to address some questions about the study. (In contrast, my thesis supervisor and I were once referred to as “ivory tower lab rats” by a commenter on The Globe and Mail, in reference to our research on optimal sexual experiences).
Commenters on the original Teach Me Tonight blog post focused on a few specific aspects of the study (I have chosen not to include the names of the commenters or the specific wording of their comments out of a desire to maintain their privacy). Some had concerns with the research question itself: Why the focus on sex and sexuality? Others expressed concerns that seminal authors in the academic romance world had been overlooked. Many expressed doubts about the size of the research sample (i.e., 20 books and 46 scenes) and the impact of this limitation on the study findings. Many others had [End Page 7] concerns about the representativeness of the sample. Why had we focused on contemporary single-title romance novels? Why had we selected RITA award-winners? In the following section, these concerns will be addressed, including the alternatives that were available, the rationale behind these decisions, and the strengths and limitations of those approaches. It should be noted that in designing the study, I prioritized my role as psychologist first, sex researcher second, and media researcher third. The intention, from the beginning, was to publish the findings in a journal devoted to sex research and so methodological decisions were made with that goal in mind.
Review of Literature and Research Question
In developing this study, an extensive review of the academic literature was conducted within psychological and social science literature. In practical terms, this review involved searches using PsycINFO (a psychology abstract database) and Google Scholar. When relevant papers were identified, the reference lists of these papers were consulted in order to locate additional sources. In addition, websites devoted to the academic study of romance novels were reviewed (including Teach Me Tonight), and Internet list serves focused on romance novel research were consulted. The website of the Romance Writers of America was reviewed, as were those of several publishing houses (e.g., Harlequin). Past studies focused on the content analysis of television, movies, lifestyle magazines, non-romance fiction and self-help books were also consulted in order to develop and refine the research methodology. Finally, publications on sexual script theory were consulted in order to develop the coding sheet and manual.
Some commenters from the Teach Me Tonight discussion expressed concern that important works in romance novel research had not been cited in the published paper. In fact, both my co-author and I had read literary criticism on the subject of romance novels by seminal authors in the field (e.g., Janice Radway, Ann Snitow, Carol Thurston, and Kristin Ramsdell); these works provided valuable information about the history and evolution of the romance novels and helped us to identify variables to evaluate. Although these sources were consulted, they were not cited in the final published paper, as the goal from the beginning had been to present the findings of this study in a psychology-oriented journal, where it is not customary that researchers in the social sciences cite literary criticism from the humanities. Given that I did not have any training in literary criticism, it did not feel reasonable or fair to critique work done outside of the social sciences. In retrospect, the wording of the final published paper was misleading: Rather than state that little research had been done to date on sex and sexuality in romance novels, the wording of the paper should have specified that little research had been done within psychology or the social sciences, whereas there has been a rich tradition of examining romance novels within literary criticism. I quite agree with the Teach Me Tonight commenters that we barely scratched the surface of the body of literature in this area.
Some commenters who participated in the discussion wondered why the focus of the study was on portrayals of sex and sexuality in romance novels. For me, it was never a question of studying anything else! My research studies to date, including my MA and PhD dissertations (Ménard, 2007, 2013), as well as my “side projects” (e.g., Ménard & [End Page 8] Kleinplatz, 2008; Ménard & Offman, 2009; Weaver et al., in press), have focused exclusively on the study of sex and sexuality, which reflects my interest and expertise as a researcher in psychology. I do agree with many of the commenters on the Teach Me Tonight discussion that romance novel research within the social sciences has been extremely limited in its focus, which may perpetuate unwarranted stereotypes about romance novels (e.g., they are “pornography for women”). This is undoubtedly frustrating for those who would prefer that researchers take a broader, more comprehensive look at romance novels. Hopefully, more research can be done on romance novels within the social sciences that will increase the visibility of the field and pave the way for other psychologists to research a broader spectrum of topics within these books—but as I say, my own interest and expertise guided this particular inquiry.
Research Design and Methodology
From the outset, the design of the sexual script study was constrained by several limitations. Limited resources in terms of time, money, and personpower had a major impact on the study design, primarily in terms of the relatively small sample size of this study. Because I was a graduate student and not a faculty member, I was not eligible to apply for a grant from the major governmental funding body in social sciences (the Social Science and Humanities Research Council [SSHRC]) or from Romance Writers of America, which is open only to “dissertation candidates who have completed all course work and qualifying exams” (this did not include myself or my co-author when the study was conceived). My original collaborator and I feared that putting together a grant application and waiting for the results would have caused such a considerable delay that it might have been impossible to complete the project within a reasonable time frame, given other academic obligations (i.e., coursework, dissertations, clinical work). Therefore, only two researchers were available to find the books (no small feat for the older novels in the sample, some of which were out of print), read them, code them, compare codings, compile the data, and run the analyses; this lack of personpower significantly restricted the number of books that could be read. For example, the coding sheet for each novel in the sample was three pages long, and the coding form for each sex scene in the book was four pages; one book with three sex scenes would involve 15 pages of coding (3 for the novel and 3×4 for the sex scenes).
In retrospect, one of the major factors impacting the design of the study was that it was conducted in academic isolation. At the time, I was not aware of other psychologists or social scientists who were currently researching romance novels and who could have provided valuable guidance, suggestions and/or feedback based on their experience. As a researcher who believes in multi-disciplinary collaboration and consultation, I consulted many professionals during the development of this project in order to ensure the reliability and validity of the findings. These professionals included an expert in psychological research design, two experts in Sexual Script Theory, an expert in statistics, and, finally, the editors and reviewers at Sexuality and Culture. In addition, the results of the original study were presented at a meeting of the Canadian Sex Research Forum in Toronto in 2010 to a group of sex researchers from a variety of academic backgrounds (e.g., sex educators, [End Page 9] nurses, psychologists, sociologists). Although these other professionals were very helpful methodologically, none had specialized knowledge about the romance novel community or had insight into specific issues or limitations inherent to romance novel research.
Study Sample Selection
The first major consideration in designing the study was selection of the study sample; this proved much more difficult than was originally anticipated. The primary goal of the study was to look at novels that had had the potential to have a significant impact on romance novel readers (e.g., books that had won acclaim and/or books that had sold many copies, although demonstrating “impact” is certainly a vexed and difficult issue). In addition, I felt strongly about studying changes in romance novel portrayals of sex and sexuality over the time; therefore, it was important to have a standardized selection criterion.
In content analysis, a randomly chosen sample is desirable in order that results be generalizeable to the larger population of messages (Neuendorf, 2002; Riffe et al., 2008). However, the use of convenience samples in content analysis may be justified under certain conditions, i.e., “when a researcher is exploring some underresearched but important area” (Riffe et al., 1998, p. 85). There is no question that portrayals of sex and sexuality in romance novels are underresearched, and the goal of publishing this study was to provide a preliminary picture rather than a definitive statement about portrayals of sex and sexuality in contemporary romance novels. Several methods of selection were considered, and ultimately, inclusion of the winners of the RITA awards for best contemporary single title romance was deemed preferential to the others. Let me briefly explain how we came to this conclusion.
My first idea during initial study design was to study the best-selling romance novels of all time; however, I soon discovered that those statistics do not exist, a fact that was subsequently confirmed by discussions with a representative from the Romance Writers of America (RWA; personal communication, 2009).
After this first idea was abandoned, my next idea was to study romance novels that had appeared on the New York Times (NYT) bestsellers list; however, this idea also had to be discarded because of practical and methodological reasons. The original idea was to include all of the romance novels that appeared on the NYT bestseller list for November and May in 1967, 1977, 1987, 1997 and 2007 (these months were randomly selected). We consulted an archive of the NYT bestseller lists that included the top-10 bestsellers for each week going back to 1950. The creator of the archive informed us that older versions of the bestseller list included only the top 10 bestsellers for the week, plus a few other books that were categorized by the NYT as “also selling” (Petersen, personal communication, 2009). The books that appeared on the NYT bestseller list were not classified by genre, so it became necessary to research roughly 80 titles per sample year, or 400 books total, in order to determine whether or not they were romance novels. However, in doing this research, it became clear that bestselling titles showed significant stability over a month. For example, in May of 1967, only 14 different novels appeared on the list for the entire month and for the entire year of 1967, only four different titles hit the number one slot (see [End Page 10] http://www.hawes.com/1967/1967.htm). Of those 14 novels, only one could hypothetically be considered a romance novel—Valley of the Dolls, by Jacqueline Susann—and even that title might not be considered a “true” romance novel. It soon became clear that it would be necessary to go through the list of bestsellers for each year included in the study sample and to research each title. However, the idea of two people researching and correctly classifying 2,600 novels (52 weeks x 10 books x 5 years) was unrealistic and unfeasible given our limited resources; in addition, it seemed very likely that significant bias might be introduced by the researchers through incorrectly excluding and/or including books in the sample. For example, it might be difficult to distinguish between a contemporary single title romance and a novel with strong romantic elements, and there are cases where authors who write novels with very strong romantic elements (i.e., Nicholas Sparks [Breznican, 2010]) have disavowed the genre, which adds to the difficulty in classifying the work. In general, the idea of using bestseller lists for creating study samples might be problematic. The selection criteria of the NYT have never been made publicly available (Diamond, 1995), which means that selection of books for inclusion on the list may reflect unknown biases. These lists may also underestimate a book’s readership and its potential influence, which may be particularly true in the case of romance novels as it has been estimated that each romance novel purchased may be read by seven people (Thurston, 1987). Finally, previous research in which the bestseller lists have been used to select a sample of books for content analysis resulted in an undesirable overrepresentation of one or two authors (c.f., Zimmerman et al., 2002).
The idea of selecting every third romance novel at the library or bookstore was considered, similarly to Diekman et al. (2000). Selection of novels at bookstores was not realistic as any money spent on the project would come out of the co-authors’ personal funds. The public libraries in two major Canadian cities were consulted, however these presented unanticipated difficulties. It was discovered that the selection of romance novels at public libraries may be biased towards or against particular sub-genres of romance novels, which may also be a limitation inherent to randomly selecting these books from bookstores. For example, one library that was considered showed a significant over-representation of paranormal romances, suggesting a preference on the part of the librarian or perhaps of the readers in that catchment area. University libraries were not useful for study sample selection as they tend to stock literary criticism (including scholarly works on romance novels, at least potentially), but, as a rule, they do not stock popular fiction of any kind. Interlibrary loans are available that allow access to many public libraries throughout North America; however, books that were ordered using this method took several weeks to several months to arrive. In addition, neither bookstores nor libraries tend to stock older romance novels, likely due to physical deterioration of the books, and so it would not have been possible to make the comparison in portrayals of sex and sexuality across time.
RITA Award Winners
In the end, then, the decision was made to study the RITA award winners for best single-title contemporary book. Presented by the RWA, RITA awards acknowledge [End Page 11] excellence in one of the categories of romantic fiction. Using the RITA award winners ensured that books within the sample were comparable to each other and were clearly defined as romance novels since the book authors themselves chose the categories that they felt were most reflective of their novels’ content. RWA’s process of allowing romance novel writers to select and rank the novels seemed democratic and valid on the surface, although some commenters on the Teach Me Tonight post suggested that this was not the case and that the awards are quite subjective. From my perspective, the RWA was extremely helpful and cooperative in answering questions about romance novels. A representative from the RWA provided extensive details about the competition process, suggesting a transparency in their approach to selecting award winners.
Clearly, RITA award-winners would not be representative of all romance novels published in any given year, or even of all contemporary single-title romance novels. However, it was thought that the sample might reflect the general trends within the industry as well as the novels that might have more of a potential to influence fellow authors, somewhat like studying popular movies by focusing on Oscar winners. In addition, virtually all of the books that were included in the sample had also appeared on one or more bestseller lists and so had the potential to influence a significant number of readers. A major limitation of using RITA award winners is, obviously, generalizeability: Results may only be generalizeable to other award-winning novels and/or to other novels written by the authors included the sample. However, it was hoped that including books published over the course of 20 years would help to address issues of generalizeability. In addition, by including books selected by community itself as “exemplary,” it was hoped that the sample would reflect a kind of distillation of the best that the community has to offer.
Considering all of the subgenres represented by the RITA awards, we chose to limit our scope to contemporary romances (as opposed to, for instance, paranormal, historical, and inspirational romances) because they would be most likely to reflect current social mores regarding sexuality. Portrayals of sexuality in other romance novel sub-genres might be influenced to a greater degree by the set of rules particular to that sub-genre rather than by social norms, which was the primary research question. For example, historical romance novels might be less likely to include depictions of safer sex behaviour or contraception if birth control was not available in the historical period during which the book is set. Sex scenes in paranormal romances might be strongly influenced by the fantastical elements of the novel rather than by social conventions. In addition, certain sub-genres have only recently become more popular, and so it would have been impossible to make a comparison across time. Mixing different sub-genres in the same sample would have introduced confounds into the results and would have necessitated a much, much larger sample. It would certainly be interesting to replicate the study using a different sample to determine the degree to which depictions of sex and sexuality differ based on sub-genres. I had originally hoped to do another study comparing contemporary to historical romance novels and retain hope that this type of study may be possible at some future date. The field [End Page 12] of psychology itself is limited by the need to ask focused and parsimonious research questions, which may limit our understanding of complex concepts in their entirety.
Single titles were chosen, as opposed to “series” novels, because it was believed that such novels would more likely reflect the perspective and beliefs of the author rather than those of the publisher. For example, Harlequin publishers specify that authors wishing to write novels for the “Blaze” series are expected to include “fully described love scenes along with a high level of fantasy, playfulness and eroticism”; heroes and heroines are expected to be in their early 20s and up (Harlequin, 2009). It was thought that the sexual content of such novels might be more homogenous than that of single-title contemporary romances.
Some commenters on the Teach Me Tonight blog post expressed concern that the researchers were not aware of the existence of sub-genres within the romance novel world. This was not true (not least because of my original co-author’s fondness for historical romances!). It was certainly never intended that conclusions from the sexual script paper be generalized to romance novels as a whole. In retrospect, the word choices that were used in the published paper may have obscured this point. Rather than use the term “romance novels,” it would have been preferable to use “RITA award-winning contemporary single-title romance novels” to clarify the limitations of the findings. Language may be an issue for those who wish to publish romance novel research, as there is often a tendency to fall back on certain heuristics to simplify communication, but which may have the added side effect of introducing confusion. (For example, one of the reviewers for Sexuality & Culture requested that references to “heroes” and “heroines” be changed to male and female protagonists.)
Study Sample Size
Many commenters on the Teach Me Tonight post suggested that the study sample (20 books) was far too small. Although this number is slightly smaller than that usually recommended for content analyses, this number of books was chosen because it was planned that the statistical analyses would be done at the level of the scene, rather than on the entire books. The original assumption was that 20 books would probably result in a sample size of 100+ sex scenes (i.e., approximately five scenes per book), rather than the actual number of 46. The limited number of sex scenes in the study sample did not become apparent until data gathering was well underway. In order to account for the limitations in the sample size, an expert in statistical analysis was consulted to determine the appropriate statistical tests to run. Despite appearances, the conclusions drawn from the statistical analyses that appear in the final paper were valid, despite the small sample size. In cases where the expected effect size is large, even small samples may produce statistically significant results, as it was in this case. So many of the sex scenes that were analyzed conformed to the expectations of traditional sexual scripts that a larger sample of RITA award-winning single title contemporary romance novels (as opposed to a larger sample that included other subgenres or categories of romance) was unlikely to generate much additional information. [End Page 13]
Certainly, it would have been preferable, as one commenter in the Teach Me Tonight discussion suggested, to read and analyze all of the finalists within a given year, i.e., eight or nine books per year, for a total of 160-180 books and potentially 400-500 sex scenes. However, a project of that size would have been tantamount to completing a second PhD dissertation and/or would have required the services of a research assistant, at the very least. In addition, it became increasingly difficult to locate the older books—i.e., those from the early 90s—as many of them were no longer in print and had to be purchased online. I agree that selecting finalists for the RITA awards would be a very viable selection method for future researchers, especially given the lack of difference that was observed in the novels over the sampling period. An expansion of the original sex script study that would include the finalists for the award might also compare the sex scenes in winning novels versus runners-up. Statistically significant differences between the two groups might suggest that adherence to sexual norms might be part of the judging criteria, whether this is a conscious or unconscious bias on the part of the judges.
Future Directions for Romance Novel Research
Romance novels are a fertile area for investigation within psychology, sociology, English, media studies, cultural studies and many other fields in academia. Some may choose to study romance novels in order to build the body of knowledge focused on romance novels; others may pick romance novels in order to study manifestations of a particular theory. I believe that either approach will build the body of knowledge and increase the legitimacy of romance novel research for others. In closing, I’d like to reflect on the challenges and opportunities facing those who pick up where my earlier study leaves off.
Challenges within romance novel research
Romance novel research is subject to a variety of unique challenges, which are likely to evolve over the next few years as the genre itself continues to evolve. Romance novels are so commonly a part of everyday life that some researchers may not deem them worthy of investigation. In particular, researchers in psychology tend to focus on negative life experiences (Seligman, 1990), while romance novels, by their very definition, are positive. The prevailing “misery mindset” in psychology represents a significant obstacle to romance novel research. It is hoped that the increasing popularity of the positive psychology movement, with its focus on positive experiences, might pave the way for more research about everyday life experiences within the field. Psychologists, in particular, would be encouraged to engage in interdisciplinary collaborations to take advantage of differing but complementary perspectives between the social sciences and the humanities.
Given that romance novels have been deemed silly, trivial and/or anti-feminist within popular conceptualizations, grant proposals based on romance novel research may be passed over by (uneducated) funding committees due to a general shortage of research funding. This situation may impose a variety of restrictions on researchers in terms of [End Page 14] procuring research materials, paying research assistants, and disseminating findings, with the result being much smaller, shorter-term investigations. The question of who is to provide the funding is complex. Over the past 10 years, the Social Sciences and Humanities Research Council has provided funding for one romance-related study, in the form of a scholarship to a Master’s level student. The RWA offers a grant of $5,000 but this amount would be insufficient to establish an ongoing, multi-year major research investigation. Nora Roberts, a prominent romance author, has endowed McDaniel College with a $100,000 grant; while this is very generous, it seems impractical for romance authors themselves to fund romance research on a large scale. Although funding from industry is common in psychology (e.g., pharmaceutical companies), it is not clear how major publishing houses would benefit from funding romance novel research. Unfortunately, this situation creates a Catch-22 whereby more research is required to increase the legitimacy of the field, but this research cannot be done without the funding that would only be accorded to an already-established field.
In some cases, the romance community itself may present challenges to romance novel researchers. Some community members may have grown tired of waging a constant battle for legitimacy and respect; the result, in these cases, may be a sense of defensiveness or suspicion regarding outsiders to the romance community. [Unwillingness to cooperate with researchers has been observed with certain minority groups (e.g., BDSM practitioners), who have felt exploited or betrayed by previous researchers (Kleinplatz & Moser, 2006)]. This may have significant implications for cooperation between researchers and the romance community. Although most of the comments on the Teach Me Tonight discussion were constructive, one commenter compared the sexual script study to spousal abuse as an example of negative attention and stated that the paper did more harm than good. Those outside the community who wish to study romance novels would be well-advised to do extensive preliminary research.
In terms of changes in the industry, e-publishing has recently taken off: Sales of romance e-books increased by 164% in 2010, with 29% of books being purchased in e-book format (RWA, 2012). In addition, there appear to be significant differences between the two groups of readers; readers who purchase e-books are younger and have higher incomes than readers of “analog” romance novels. This discrepancy may create some difficulties in determining what constitutes a representative sample: Researchers may need to decide whether to include e-books as well as print copies in their sample, or whether it is preferable to focus on just one format. The composition of the study sample may also be influenced by differences in sub-genre sales between the different formats; e-book readers, for instance, are more likely to purchase erotic and paranormal titles. Another major change within the industry, related to the increase in e-book sales, is the increasing proliferation of ever more specific romance novel sub-genres, a kind of narrowcasting to romance fans. Again, this may have a significant impact on sample composition. It may no longer be possible or accurate to draw conclusions about romance novels as a whole but only about increasingly specific sub-genres. The findings from future romance novel research may demonstrate great specificity but limited generalizeability. An alternative approach for researchers would be to focus their questions on qualities that are common across all romance novels and are therefore unaffected by the specific “rules” of the sub-genre (e.g., characteristics of the happy ending, characteristics of the central romantic relationship [RWA, 2012]). [End Page 15]
On a purely physical level, it may become more and more difficult to study older romance novels. In doing the sexual script study, it was difficult to locate romance novels from 20 years ago as many were out of print. Books that could be located on eBay or from used book stores were sometimes in poor condition due to age and poor paper quality. The inability to locate books may limit the ability of researchers to study changes in romance novels across time. I believe that studying changes across time is important to determine to what degree romance novels both shape and reflect realities for readers. Hopefully, romance novel publishers might be encouraged to continue the process of digitizing (and, thus, preserving) older novels, with the side benefit of being able to sell them as e-books. Indeed, Harlequin has already begun to take the lead in this effort, and perhaps other publishers will follow suit.”
Opportunities within romance novel research
Romance novels offer many unique opportunities to study human experiences; there are many, many aspects of these books that remain unexplored, at least in the social sciences. Although the focus of research questions has often been on representations of sex, sexuality and gender, researchers might also look at portrayals of religion, politics, family, children, physical and mental illness, employment, leisure activities, communication styles, interpersonal relationships, etc. Previous research has tended to analyze the books in their entirety, but research questions could also be focused at the level of the characters, the relationships, families, scenes, titles, or covers. Studies could be done specifically on characters’ behaviours, speech and/or attitudes. Researchers who look at questions from multiple angles would be in a position to evaluate the consistency (or not) of certain ideas within and across books and the implications of any discrepancies.
The interdisciplinary possibilities within romance novel research may present both challenges and opportunities. In academic research, the guiding values of one’s field will determine appropriate research questions, study design, the nature and process of accumulating supporting evidence and interpretation of findings. Questions about generalizeability, reliability, and validity as they relate to research methodology—questions that are paramount in psychological research—may be meaningless for literary critics. Although both social sciences and humanities are concerned with the human condition, a major difference between them is the empirical mindset that guides research in the social sciences, which suggests that gathering evidence in the form of multiple investigations may allow us to converge on some central “Truth” about the area of interest. Literary criticism, which is not guided by empiricism, allows for and encourages the validity of multiple perspectives. This fundamental difference has important repercussions for study design. For example, the need to choose a random sample of novels so that results may be generalizeable to other books is a concept that may seem absurd to literary critics. Another major distinction may be sample size; researchers from the social sciences might include more novels in their sample (relative to literary critics) but go into less depth in their analyses. This procedure may also have implications for the process of the research, as those in the social sciences are more likely to conduct their research in teams rather than as individuals. A researcher studying novels from the perspective of literature may be [End Page 16] more concerned about the artistry or ideas of a particular author, whereas a psychologist might be more interested in the impact of reading popular romance fiction on the romantic and sexual behaviors of readers.
However, there may be much to gain from interdisciplinary collaborations as they apply to romance novels. The opportunity to share theories and concepts across disciplines may help address lacunae in our own field (e.g., psychoanalytic analysis, feminist approaches). Certain research methodologies may also be complementary: A study of the metrics and rhythm in the language of romance novels would be a perfect fit with a psychological study involving discourse analysis. In our case, the focus of the sexual script study was on the manifest behaviours of the characters; a partnership with a humanities researcher might have helped to uncover additional information represented within the latent content of the novels. In general, results from the social sciences may provide a greater breadth of information about romance novels, whereas studies in the humanities may provide greater depth and complexity of interpretation.
Previous studies in the social sciences have begun to elucidate depictions of sex, sexuality and gender in romance novels. Sample size, selection criteria for novels, and inclusion of different sub-genres varied considerably across these investigations. These methodological differences have shed light on certain aspects of the books but have also created gaps in the research knowledge, suggesting further areas for exploration. The sexual script study was constrained by limitations (e.g., sample size, limited focus) but offered a broader look at sexual behaviours in the books than had been studied previously. Future researchers in the area would be encouraged to take into consideration changes within the romance industry, such as the proliferation of specific sub-genres and the rise in popularity of e-books. They might also expand their questions of interest and study the books from different angles (e.g., characters, scenes, titles). Interdisciplinary collaborations might help to provide a richer, more complete picture of romance novels.
 I hypothesize that establishing a permanent, ongoing “romance novel research laboratory” within a psychology department that would be capable of doing large-scale, multi-year studies would necessitate an annual budget of approximately $25,000 (CDN). This figure would cover research materials, office equipment and supplies, travel costs, and personnel. Personnel, in the form of graduate students and/or research assistants, would be a major expense given the time-intensive nature of this type of research.
Breznican, A. (2010, March 12). Nicholas Sparks, Miley Cyrus share a ‘Last Song’ love story. USA Today. Retrieved April 2, 2012 from http://www.usatoday.com/life/movies/news/2010-03-11-lastsong11_CV_N.htm
Brown, J.D., L’Engle, K.L., Pardun, C.J., Guo, G., Kenneavy, K., & Jackson, C. (2006). Sexy media matter: Exposure to sexual content in music, movies, television, and magazines predicts Black and White adolescents’ sexual behavior. Pediatrics, 117, 1018-1027.
Clawson, L. (2005). Cowboys and schoolteachers: Gender in romance novels, secular and Christian. Sociological perspectives, 48, 461-479.
Cox, A., & Fisher, M. (2009). The Texas billionaire’s pregnant bride: An evolutionary interpretation of romance fiction titles. Journal of social, evolutionary, and cultural psychology, 3, 386-401.
Diamond, E. (1995). Behind the Times: Inside the New York Times. Chicago, IL: University of Chicago Press.
Diekman, A.B., McDonald, M., & Gardner, W.L. (2000). Love means never having to be careful: The relationship between reading romance novels and safer sex behavior. Psychology of Women Quarterly, 24, 179-188.
Fisher, M., & Cox, A. (2010). Man change thyself: Hero versus heroine development in Harlequin romance novels. Journal of social, evolutionary, and cultural psychology, 4, 305-316.
Gagnon, J. H. (1977). Human sexualities. Glenview, Il: Scott, Foresman and Company.
Gagnon, J., & Simon, W. (1973). Sexual conduct: The social sources of human sexuality. Chicago, Il: Aldine Publishing Company.
Harlequin (2009). Writing guidelines: Harlequin Blaze. Retrieved March 12, 2009 from http://www.eharlequin.com/articlepage.html?articleId=544&chapter=0
Kleinplatz, P.J., & Ménard, A. D. (2007). Building blocks towards optimal sexuality: Constructing a conceptual model. The Family Journal, 15, 72-78.
Kleinplatz, P.J., Ménard, A.D., Paquet, M.-C., Paradis, N., Campbell, M., Zuccarini, D., & Mehak, L. (2009a). The components of optimal sexuality: A portrait of ‘great sex’. Canadian Journal of Human Sexuality, 18, 1-13.
Kleinplatz, P.J., Ménard, A.D., Paradis, N., Campbell, M., Dalgleish, T., Segovia, A., & Davis, K. (2009b). From closet to reality: Optimal sexuality among the elderly. The Irish Psychiatrist, 10, 15-18.
Kleinplatz, P.K., & Moser, C., (2006). Sadomasochism: Powerful pleasures. Binghamton, NY: Harrington Park Press.
Lacy, S., & Riffe, D. (1996). Sampling error and selecting intercoder reliability samples for nominal content categories: Sins of omission and commission in mass communication quantitative research. Journalism & Mass Communication Quarterly, 73, 969-973.
Ménard, A.D., & Cabrera, C. (2011). ‘Whatever the approach, Tab B still fits into Slot A’ : Twenty years of sex scripts in romance novels. Sexuality & Culture, 15(3), 240-255.
Ménard, A.D., & Kleinplatz, P.K. (2008). 21 moves guaranteed to make his thighs go up in flames: Depictions of “great sex” in popular magazines. Sexuality & Culture, 12, 1-20. [End Page 18]
Ruggiero, J.A., & Weston, L.C. (1983). Conflicting images of women in romance novels. International journal of women’s studies, 6, 18-25.
Romance Writers of America [RWA], (2012). About the romance genre. Retrieved March 27, 2012 from http://www.rwa.org/cs/the_romance_genre/romance_literature_statistics/industry_statistics
Seligman, M. E. (1990). Learned optimism. New York: Alfred A. Knopf, Inc.
Simon, W., & Gagnon, J. (1986). Sexual scripts: Permanence and change. Archives of sexual behavior, 15, 97-120.
Simon, W., & Gagnon, J. (1987). The sexual scripting of oral genital contacts. Archives of Sexual Behavior, 16, 1-25.
Thurston, C. (1987). The romance revolution: Erotic novels for women and the quest for a new sexual identity. Chicago, IL: University of Illinois Press.
Zimmerman, T.S., Holm, K.E., Daniels, K.C., & Haddock, S.A. (2002). Barriers and bridges to intimacy and mutuality: A critical review of sexual advice found in self-help bestsellers. Contemporary family therapy, 24, 289-311. [End Page 19]