The World of Teenage Television: A Longitudinal Content Analysis of
Prime-Time Programming Most Popular Among Adolescents

Amy Richards Franzini
Widener University

   Whether it is Elvis gyrating on The Ed Sullivan Show or Marissa overdosing on The O.C., television shows that are popular among adolescents often create parental and/or public concern. Just what ARE our adolescents watching on television? This study attempts to answer that question, by content-analyzing a full season of television programs most popular among adolescents.

   Nielsen estimates that at least a million adolescents view television nightly (Johnson 1998). According to a recent study done by the Kaiser Family Foundation, adolescents spend an average of three hours each day watching television (2005). There is little argument that adolescents are watching television content; the argument begins when one discusses the influence or importance of that content.

   While direct influence is impossible to prove, there are many theories that suggest that television is an important (if not the most important) socializing agent for adolescents.

   Theories of Socialization Through Television

   Bandura’s (1986) Social Cognitive Theory suggests that people may identify with characters on television and develop expectations about what it means to be someone who is “like” that character. Therefore, adolescents would develop expectations about what it means to be a teenager through the teen characters they watch on television.

   Gerbner’s (1976) Cultivation Theory takes a more long-term, cumulative approach to socialization. Cultivation Theory attends to the common themes that intersect all the television programs a person grows up watching. Consequently, if a person sees certain values and norms over-and-over again, that person is more likely to “cultivate” similar values.

   Other theorists take this socialization a step further, positing something called wishful identification, a psychological process in which an individual wants to - perhaps even tries to - become like a television character (v. Feilitzen & Linne, 1975; Hoffner, 1996; Hoffner & Buchanan, 2005).

   Both Parsons (1951) and Geertz (1973) recognized the power of socialization and enculturation to show members of a culture not only codes of behavior, but also the norms and values of that culture. Television offers a lifelong inculcation of a culture’s values and norms, from toddlerhood to adulthood, but it can be especially interesting to look at those values and norms portrayed during that time in a person’s life when they are developing their own personal value system – adolescence. This study, then, sets out to better understand the general landscape of the world of adolescent television, through a season-long longitudinal content analysis of ten television programs most popular among adolescents: Who makes up this world and what are the norms and behaviors of its “culture”?

LITERATURE REVIEW

   The sociology-of-adolescence literature generally stresses how various social dynamics create adolescent values, norms and expectations, offering a wealth of data relating to such factors as religion, geography, race, ethnicity, intelligence, academic achievement, family background, parental support and control, and peer-groups influence (see, for example, Coles and Stokes, 1985; Hayes, 1987; Lees, 1993; Moore & Rosenthal, 1993; Irvine 1994). It is also almost uniformly recognized that in addition to these factors, the media play an integral role in adolescent socialization. While this literature is sometimes broad-based -- regarding all media collectively as influential in the adolescent-maturation process (e.g., Arnette 1995, Steele and Brown, 1995; Strasburger, 1995), most of the work tends to be medium-specific. Research has focused on adolescents and film (for example Charters, 1933; Hoffner, 1995; Johnston, 1995; and Oliver, 1993): adolescents and radio (such as Brown, Childers, Bauman & Kauch, 1990): adolescents and video games (for example Funk and Buchman, 1995 & 1996; Irwin and Gross, 1995; Funk and Buchman, 1996; and Dill and Dill, 1998); adolescents and magazines (such as Pierce, 1990, 1993, 1995; and Garner, Sterk and Adams, 1998); and adolescents and popular music (for example Roe, 1990; Christenson and Roberts 1990; Arnett, 2002; Christenson & Roberts, 1998).

   Still and all, probably the most considerable amount of the “media and adolescence” literature refers to television, which Elkin and Handel (1978) regard as that medium most influential on youth.

   There has also been much research into the general socializing function of television in the lives of adolescents, arguing that television is a powerful socializing force because of its ubiquitousness and its undemanding nature in terms of required interpretational skills (Schramm, Lyle and Parker, 1961; Berry, 1980; Roberts and Bachen, 1981; Berry and Mitchell-Kerner, 1982; Roberts and Moccoby, 1985; Dorr, 1986; Wright, 1986; Hawkins Pingree & Adler, 1987; Liebert and Sprafkin, 1988).

   Among the various areas of socialization to which this literature attends is that of moral instruction. Researchers have explored how television contributes to the ethical development of adolescents and teens (Meyer, 1976; Ryan, 1976; Sutherland and Siniawsky, 1982; Kohlberg, 1984; Lindz, 1984; Bryant and Rockwell, 1994; Tan, Nelson, Dong and Tan, 1997; Kremar and Vlakenberg, 1999; Rosenkoetter, 2001). By and large this scholarship argues that there is an inverse relationship between television viewing and high moral standards. However, Messaris and Sarett (1981) found that television actually provides increased opportunity for moral instruction by presenting moral situations that parents and children may discuss together.

   Besides research into the general socializing function of television, there has been much research relating to specific content areas relevant to adolescence, such as sex or violence.

   Strasburger and Wilson (2002) provide a summary of the hundreds, if not thousands of scholarly articles dealing with the potential effects of television violence on young people. In the main, these articles suggest that television (1) facilitates aggressive or anti-social behavior, (2) desensitizes youth to “real-life” violence; and (3) increases fears of a dangerous world.

   Clearly, given that sexual development is central to our understanding of adolescence, much of the literature relating adolescents and teenagers to the media, at least in part, also relates to sex. This research typically either relates to the actual sexual content of programming (for instance Ward (1995; Greenberg, Stanley, Siemicki, Heeter, Soderman, and Linsangan, 1993, as well as Cope-Farrar and Kunkel, 2002) or the correlation between viewing that content and adolescent behavior (such as Greenberg, Brown & Buerkel-Rothfuss,1993; Brown and Newcomer, 1991; and Peterson, Moore and Furstenberg, 1991).

   Still, there lacks in this body of research, a more general content analysis of that programming most popular among adolescents. While some content analyses focus specifically on teen programming (Ward 1995 and Greenberg, Brown & Buerkel-Rothfuss, 1993) they focus strictly on sexual content. It is important, then, to study the programs most-viewed by adolescents to gain a greater understanding of who inhabits this world and the general values and norms, by content analyzing other areas as well, such as academics, alcohol-use, family interaction and friendships. This study, then, provides such an analysis.

METHOD

   In order to study the landscape of teen television, a sample of programs was chosen and a coding instrument constructed to best analyze the content.

   All episodes (excluding reruns) of an entire season (1998-1999) of prime-time, network shows most viewed by adolescents and teenagers (12-17 year olds) were studied: Beverly Hills, 90210; Boy Meets World; Dawson’s Creek; ER; Friends; Home Improvement; King of the Hill; Party of Five; Sabrina, The Teenage Witch; and The Simpsons.

   The unit of analysis for the study was the individual character. More specifically, the study analyzed all plot-functional characters who were adolescent or older, i.e., all characters who might likely serve as a social model for the adolescent audience. A plot-functional character was one playing an integral role in the narrative of the show. This character-as-unit-of-analysis was chosen to provide a clear route to character (and therefore program) values, norms and behaviors. If, in socialization, adolescents model their behavior after others, it is appropriate then, to study not simply behaviors, but rather the character performing such behaviors. This character-focus also allows for coding of topics difficult to code from a behavior-standpoint, such as popularity and family-life.

   The coding instrument individually coded each character in each episode of his/her appearance and contained 47 variables organized around nine categories: (1) Programming (including seven variables), (2) Demography (including five variables), (3) Appearance (including three variables), (4) Moral Qualities (including eight variables), (5) Academic Success and Intelligence (including five variables), (6) Professional Performance (including five variables), (7) Family Relations (including three variables, (8) Personal Success (including six variables), and lastly, (9) Sex & Romance (including 5 variables distributed among three subcategories).

   The data emanating from the episodic coding instrument were used to compile a seasonal history for each core character—as opposed to an account of activities isolated to single episodes. What this basically entailed was a meta-analysis of 44 of the original variables (all but the programming variables), whereby a synthesis of the character’s behavior could be constructed.

   This meta-analysis consisted of doing a few different things so as to provide an historical view of how characters were represented. For all variables but those pertaining to moral qualities as well as sexual and romantic behavior, a seasonal (core-character) value was assigned by selecting the mode – that value coded most (across all episodes) for the given character. In other words, if a character was assessed as likeable in 20 of 25 episodes, that plurality became the core-character value, i.e. the character was, therefore historically coded as “likeable.”

   With respect to the moral qualities, all of which were coded on a binary presence/absence basis, each character was historically characterized again by presence or absence -- only this time on a seasonal, rather than episodic basis. That is, if a character was coded in one or more episodes as possessing or enacting the given “moral quality” variable, his/her seasonal, core-character assessment indicated presence or possession. This binary coding system was also used for the psychotherapy variable.

   Finally, with respect to sexual variables, longitudinal assessments were constructed by employing an essentially cumulative technique with the relevant variables. More specifically, the highest level of sexual intimacy and the number of partnerships into which each character entered was ultimately collapsed into one sexual history variable.

   All 231 episodes of the 10 shows were videotaped for subsequent analysis by both the researcher, who served as primary coder, as well as another trained coder, who provided reliability data. Reliability was determined on the coding of a random 10% of episodes, requiring 85% agreement on each item^[i]. On average over 90% reliability was obtained on all items. Specific, intercoder reliability was not conducted with respect to the longitudinal, core-character data because they were all meta-analytic, i.e. resulting from techniques that collapsed the episodic data already checked for reliability.

RESULTS AND DISCUSSION

   Collectively, the ten programs sampled featured 144 adolescent-or-older, plot-functional characters (who made a total of 1656 appearances in the 231 episodes). As described, for every episode in this sample, each plot-functional character was coded in terms of four broad categories: (1) Demography (five variables), (2) Appearance (three variables), (3) Social Indices (28 variables), (4) Sexual and Romantic Behavior (six variables)^[ii].

Character Demography

   Approximately 54% of the core-characters were male and 45% female. In regards to race, Caucasian characters dominated the landscape (at over 90%). African-Americans constituted around 5%, Asian-Americans about 2%, and all other races combined, fewer than 2%.

   Characters’ ages were coded as follows: teens, young adults, middle-aged characters and seniors. Approximately one-third of the characters were young-adults, another third were middle-aged, and the remaining third was constituted by teens and seniors – with teens representing approximately one-fourth of the total group; and thus, very few much-older characters.

   In regard to characters’ social class, the majority of characters were middle-class -- around 85%. There were not many upper-class characters (a little over 10%) and even fewer working-class characters (close to 5%).

   The sample was overwhelmingly heterosexual (95% of characters). Almost 4% of the core-character sample was homosexual, and the remaining 1% was unclear.

   It may be of interest to discover if the demography of the shows to which teenagers are reportedly attracted differs from the demographic landscape of fictional television in general. In making such comparison, one caveat should be noted at the outset. The sampling technique for this study was different from those producing the data with which it is to be compared. The demographics in the extant literature, although also available in terms of plot-functional characters were typically based on a sample of a composite week of television programming. As seen, the data from this study emanate from a sample of the standing core characters in the sampled programs.

   To begin with, these shows most viewed by teenagers provided, in most instances, a more “liberal” demography than that provided on television programming in general. While about 45% of the characters of this study were female, general television statistics show a female sample of 37% in a composite week (Gerbner, 1998). Similarly, in considering the representation of social class, the present sample revealed a greater representation of working class characters than generally available in television programs – 4.9% - versus 1.4% in Gerber’s composite-week sample (1998). Also, there was a higher percentage of homosexual core characters (3.5%) than that found in the general population of television shows, 1%-2% (Seomin, 1999). In terms of race, however, the shows sampled presented fewer African-Americans than found in the general TV population: 5.6% of core characters versus 12.3% in the composite week (Gerbner, 1998). Conversely, these shows provided a comparatively greater number of young-adult characters, 35.4% versus Gerbner’s (1998) 32.2%.  

   The question then remains as to what, if anything, would account for a seemingly more liberal representation in terms of gender, social class, sexual preference and age but not race. That is to say, as noted above, the staple characters populating programming most viewed by teenagers include as many or more women, working class characters, and young people then television typically offers, but fewer racial minorities.

   From an armchair perspective, it would be too rash to infer some special institutional ideology (in this case, racism) on the basis of these relatively simple findings. Perhaps, a more likely hypothesis might be that the demography reflects that which is assumed resonates with the consumer audience to whom these shows are presumably aimed, to wit, younger, Caucasian females. Still, the proportionately greater representation of working class characters in these shows may not be explained by this “commercial” interpretation. Clearly a more intensive and expanded analysis of audience demographics as well as, perhaps, institutional investigation of the shows’ producers would be necessary before a more conclusive understanding is derived.

Appearance

   Appearance involved three variables: style of dress, weight, and overall looks. Core-character measurements relied on finding the mode condition among the characters’ total-appearances and using that central tendency to characterize the player more generally, For example, if a character was coded as being good-looking in 17 of 25 appearances, his or her core-character value was considered good-looking.

   A character’s style of dress was coded as: sexy (tight clothes, cleavage, etc.), average (all dress styles that were neither sexy nor prim), or unsexy (notably non-sexual, ultra-conservative, prim, matronly, etc). Core-character’s dress was mostly assessed as average (94.4%). There were few characters whose dress was considered sexy (fewer than 5%), and even fewer characters whose dress fit the unsexy condition (less than 1%). Thus, although there seems to be considerable furor (from, for instance, Parent’s Television Council and National Organization of Women) over the provocative fashions of teenagers’ pop- icons (e.g. Britney Spears, Christina Aguilera, and so forth), it cannot be said that those stylings were much offered in these television programs most watched by adolescents and teenagers.

   The weight variable indexed simply whether or not a character was considered overweight. Core-characters were typically not overweight (95%). The meaning of these data are somewhat compromised by the coding instrument, which, in an effort to obtain reliable data, sacrificed specificity, given that earlier studies coding characters’ weight and body types have had significant reliability problems when attempting to make fine distinctions, (e.g. even for differentiating “thin” from “average”). This makes sense to the extent that if, for example, characters portrayed by Calista Flockhart or Lara Flynn Boyle would necessarily have to be coded as “thin”, then it apparently becomes difficult to make more precise gradations within the rest of the character population. This coding handicap is relevant here insofar as this study did not differentiate particularly thin or skinny characters from thin characters, or then again moderately round individuals from obese ones, and such differentiations might have otherwise been deemed relevant with respect to teenagers and body image. On the one hand, this study is able to suggest that those programs most watched by teenagers don’t much show characters who would likely be socially defined as having weight problems. This, in turn, might be translated by inclined activists to mean that a more normal distribution of body types (i.e. normal in the real-life population) are not made available to viewers, and may, therefore, provide a source of tension vis a vis body image. On the other hand, the absence in this study of measurements that might indicate extreme thinness prevents these data from being used to suggest such things as the promotion of eating disorders (as suggested in, for example, Hofschire & Greenberg, 2002; Brumberg, 1997; and Fallon 1990).

   There are a few points that should be specifically discussed with respect to overall looks. For one, it should be remembered that the coding on this variable was based on authorial intent – how the character seemed to be “scripted” or was to be seen by other characters. This technique of coding relatively subjective variables (like appearance) has the advantage of characterizing actors (and thus stories and their morals) as they are “built into” the drama; people are coded as their character. This allows, for example, Sarah Jessica Parker to be a plain, unsexy, geekish-type (Square Pegs) as well as a good-looking, sexy, provocateure (Sex and the City). It also allows for important distinctions that might not, using a different coding technique, be observed. That is to say, as a lot, major actors and especially actresses, are good-looking, they can be dressed down for certain roles – sometimes even making real physical change (i.e. roles held by Robert DeNiro, Tom Hanks, and Renee Zellweger in Raging Bull, Cast Away and Bridget Jones’ Diary, respectively). Were an alternative coding technique employed in coding appearance variables, such as using everyday life as the comparative standard, most all major actors would necessarily be coded, at the very least, as good-looking, even if, in the context of the drama, that isn’t at all how they are “supposed” to be assessed.

   Over 40% of the core sample was coded as good-looking. Few characters were rated unattractive (fewer than 4%). Over half of all core characters were rated average.

   Given the above description and rationale for the coding technique, it is more understandable, then, to discover that over 50% of the core characters in this sample were coded “just” as average. However, even with that caveat -- that is, that it is unlikely that most of the average actors are really average by everyday standards -- it is interesting to note that the proportions change when considering the sum total of appearances: the good-looking characters made proportionally more appearances than the average ones (almost 60% of total appearances for good-looking characters versus almost 40% of total appearances for average characters). The fact that good-looking characters were “over-represented” then, attends to an issue that certain activists would consider serious when related to teen-watched shows, to wit, -- self-image. It has been much argued (Hofschire & Greenberg, 2002; Brumberg, 1997; and Fallon, 1990; a full compendium of resources on teens and body- image is available at www.about-face.org) that the pressures of adolescence are great with respect to appearance; thus an “over-abundance” of especially good-looking characters might be thought to intimidate psychologically these younger viewers.

   Of course, there is counterpoint to this line of thought: especially attractive performers are, in fact, cast in movies and television shows so as to attract (in this case) the teenage audience; and it is simultaneously assumed that teenagers choose to view those media endeavors, especially these particularly attractive performers. The question, then, might be: would these younger people ever choose programs featuring more regular-looking characters over those featuring especially good-looking ones? Would teenagers just as easily develop crushes (an obvious motivation in media selection) on average-looking characters?

Moral Qualities

   Eight different “moral” variables were coded: cigarette smoking, alcohol use, drug use, criminal activity, duplicitousness, religious activity, Samaritanism, and civic service. Each of these variables was subject to binary assessment in terms of whether or not they had ever (even once) engaged in the given activity during the season.

   Four of the 144 core-characters smoked at least one cigarette during the season. Almost 40% of core characters consumed alcohol at least one time during the season. When alcohol consumption is correlated with age, over 80% of the core characters consuming alcohol were young adults or older. One character (a young adult) used drugs (in 4 episodes) during the season.

   It is a judgment call to say categorically what four incidents of drug use distributed among 244 different television episodes throughout a season represent. However, it is relatively safe to say that, compared to the other activities this study has indexed (e.g. alcohol consumption, dating, religious activity, etc.), illegal drug use wasn’t represented very much -- and the one character who was represented as an illegal drug user was not a teenager.

   Often, the problem with interpreting these sorts of data is that the media can be “damned if they do and damned if they don’t”. For instance, in terms of the data provided by this study, one can hypothetically hear cries that these shows so popular among teenagers are indifferent to the youth/drug problem because “so little” is represented. Conversely, were all these shows to feature some representation of teenage drug experimentation -- even possibly with very negative consequences -- one can also hypothetically hear cries that television glamorizes drug use, particularly if the users are physically appealing teenagers.

   Any illegal behavior (besides that connected to drug or alcohol use) was recorded, whether or not the character was caught and/or charged with a crime. Almost 10% of characters committed some sort of crime, overall there was criminal activity in only 1.1% of all character appearances. Thus, it can be said that there were no sustained character portrayals of criminal activity in these shows most watched by teens.

   Duplicitousness encompassed outright lies, misrepresentation, malicious gossip, backstabbing and confidence betrayal. Almost 20% of the core characters behaved duplicitously at least once during the season; however, as a whole, characters were duplicitous in fewer than 5% of all their appearances. Therefore, it can be said that relatively few characters were consistently duplicitous throughout the season, but on occasion, characters behaved duplicitously.

   The occurrence of any religious activity by a character, such as praying or going to church, was recorded. While about 15% of core characters participated in such religious activities, fewer than 2% of total appearances actually featured church-going and so forth. Characters may have said a prayer or attended a funeral once during the season, but there was no sustained portrayal of religion in the sample.

   A little over 10% of core characters were, at some point, Samaritans (demonstrating of sincere concern for relative strangers (i.e. not friends or family). However, as with religion, Samaritanism was present in fewer than 2% of all character appearances. Five characters (3.5%) performed a total of nine charitable or civic acts (0.5%) during the season. Again, there was no maintained depiction of charity or civic service in the sample.

   It is common, but problematic, to assign values of relative magnitude to percentage occurrences where no logical norm or standard really exists. That is, it is common to reason that a single-digit percentage value represents an uncommon phenomenon, whereas a higher, double-digit percentage, conversely, indicates a common happening. While, certainly, these might be statistically reasonable inferences, this same logic might not apply to a proper contextual understanding of the moral data just presented.

   From one perspective, it would be easy to say that there was a “low” incidence of smoking, drug use, criminal behavior, or the arguably more positive moral qualities of religious activity, Samaritanism or civic service because in all cases, the proportion of per-appearance occurrences fell below 2%. Indeed, it might also be interpreted that even alcohol consumption was “low” because “fewer than” 20% of all appearances included drinking. However, it might be problematic to invoke terms like “high” or “low” simply on the basis of the numbers without considering the norms and patterns of narrative structure and dramatization. For instance, smoking is now rarely dramatized on television. Indeed, the most likely place to see smokers may be in public-service announcements warning against it. Considering this, even 28 cases of tobacco use might be seen as relatively considerable. Of course, this is not to say that what is numerically low is not also “conceptually” low. For example, that characters behaved duplicitously in 4.3% of their appearances might, indeed, be unusually low given that several of the programs in the sample were hour-long semi-serialized dramas – a genre that is often driven by “villainous” characters’ scheming and morally objectionable behavior. Unfortunately, the type of comparative data required to make accurate inferences of this nature are not available with respect to this study.

   Thus, all that can be said here is that there was a relatively greater incidence of alcohol consumption by characters than the other morally-questionable activities measured. However, as already noted, when alcohol use was cross-tabulated with age, we found that about 75% of that alcohol use was, in fact, limited to adult usage.

   When examining the remaining “prosocial” moral indices, it might likewise be observed (again, on a statistical basis) that their incidence was low. However, without a previously-described, comparative frame of reference, it is difficult to argue their signification.

Academic Success and Intelligence

   Characters were coded for their level of education, academic performance and whether or not they were the recipients of praise and/or criticism for any academic activities. Additionally, characters were also assigned an overall intelligence rating.

   The highest level of education a character reached could be coded as current high school student, high school graduate only, current college student, college graduate only, and graduate or professional school. Of the core characters, almost 40% were high school graduates (only). Characters having graduated college (only) constituted almost 25% of the core-character sample, while a little over 15% of the core characters were currently in high school.

   When applicable, the academic success of characters (in school) was indexed. For much of the sample, this category was either not applicable or not indicated, typically because characters were no longer in school, or if they were, their academic performance was not shown.

   When characters were shown in the classroom during an episode, their academic accomplishments were mostly average or nondescript. In approximately 2% of the sample, characters were coded highly successful in academics; and another 2% were above-average. Characters were coded as academically unsuccessful in fewer than 2% of the sample.

   About 3% of core characters were criticized for their academic performance, and almost 5% of core characters were praised for their academic performance at least once during the season. Still, academic praise or criticism was present in fewer than 1% of total appearances.

   Characters were generally assessed as possessing average intelligence (about two-thirds of core characters). The average value accounted for characters who were neither particularly smart nor substantially dumb, as well as characters for whom an intelligence rating was not available.

   Perhaps it might be argued that “average” and “not available” are not the same thing and should appear as separate values. However, in all cases, there was an attempt to streamline the construction of variables so as to determine the clearest and most significant patterns) and in this case (and in other variables to be examined that mimic the same logical structure) the “average” value is used to code both characters who are assessed as not having extreme (or remarkable) attributes or behavior in either direction or for characters whose screen-time/performance does not provide an opportunity for such assessments to be made (i.e. normally, “n/a” values). The logic behind combining these two values -- although analytically separate -- is as follows: when considering narrative structure, authorial intent, or any other term that may be used to identify what the story is seemingly representing, the “average” and “n/a” values indicate that the story is not defining or drawing attention to a character in terms of some non-ordinary quality. In this sense, then -- especially if the subsequent analysis concentrates on representations of the more extreme qualities -- the “average” and “n/a” values are functionally quite similar. Clearly, in terms of variables where either “average” and/or the particular absence of an event has theoretical significance, in and of itself, collapsing the two would not be appropriate.

   Keeping that coding caveat in mind, then, again, in about two-thirds of the core-character and sample, characters’ intelligence was typically measured as average. About 20% of characters were coded as above-average -- moderately learned and socially capable. Fewer characters were coded as street smart (3.5%), intellectually erudite (1.4%), or dumb (4.2%), with the fewest being erudite. Generally, the characters seen on the television shows most viewed by teenagers were neither remarkably bright nor noticeably dimwitted; in fact, their intelligence was rather unremarkably and unnoticeably average.

   As the foregoing data pertaining to the representation of academic issues tell us, approximately one-quarter of the core-character sample was constituted by those attending some sort of educational institution; however, relatively speaking, the academic performance of these characters was infrequently indicated in any way. All in all, and again, the relative inability to code various indices of academic performance would seem to suggest that -- although not measured directly -- characters in these shows were not shown in a classroom on a regular basis. Also, the overall intelligence of characters both in and out of school was most typically average. Since a substantial portion of teenagers’ time is spent in the classroom, it may be looked at as odd that the shows they typically view ignore this aspect of their lives; however, from a uses-and-gratifications perspective, these shows could also be viewed to have an “escapist” quality to them -- allowing teens to forget about their everyday lives and escape into a world where schoolwork is not so ever-present and important…or, it may just be, perhaps, that this belief dominates the mindset of the producers of such shows. In this light, it would be interesting to collect the Nielsen demographics for shows whose major focus was high school, from Room 222 through Fame to Boston Public.

Professional Performance

   Professional Performance indexed a character’s role in the workplace in terms of three variables: extent of employment, professional category and professional success.

   The extent of employment of a character was coded as: full-time, part-time, non-employed (by choice – homemaker, student, independently wealthy, etc.), or unemployed (not by choice).

   About half of all core characters were employed full-time and approximately 10% were employed part-time. Fewer than 2% of all characters were unemployed, and almost a third of all characters were non-employed (by choice). At first glance, given that a third of the characters were non-employed, one might conclude that these characters made a conscious choice not to work, which might be argued as sending an inappropriate message to teen viewers. However, it must be remembered that most of the non-employed characters were students. In retrospect, to distinguish properly between student characters and those non-student characters voluntarily unemployed, separate values -- non-employed and student -- should have been included.

   In coding characters’ professional category, work information was not identified with respect to about 40% of characters. Considering all codable and uncodable portrayals, the medical field was the most represented (12.5% of core-characters). The next highest represented professions included media (8.3%), service-industry (8.3%) ”other“ white collar (6.3%), blue-collar (5.6%) and education (5.6%). It is interesting to note that law enforcement, which is generally considered to be over-represented on television, only accounted for 2.8% of core-characters on programming most-watched by adolescents.

   Characters’ professional success could not be coded for over 60% of the core-characters. Considering this, about one-third of the sample included working characters with nondescript or un-noteworthy employment situations, while approximately 5% were coded as successful, and fewer than 2% were coded as unsuccessful. Just as student characters weren’t much dramatized in the classroom, workers don’t seem to be much dramatized at their jobs; and even when the workplace is represented, it is often background or commonplace.

   In these shows most popular among youth, about half of all characters were shown as having full-time employment, though it must be remembered that several programs concentrated on non-working student characters. Still, these shows did not seem to provide too much characterization of professional life, even for employed characters. While this may seem to be an artifact of having a number of programs centrally focusing on not-yet-employed youth, it must also be remembered that a similar proportion of student characters were not shown in academic settings either.

   Again, the extent to which these particular dramatic representations are typical requires a larger and unavailable set of data for comparison. It may actually be the authorial intent or dramatic convention of shows not focused in the workplace (or school) to underplay these areas, while concentrating on other dramatic areas, such as family, friendship, or romantic relationships, which may be thought richer fields for dramatic mining.

Family Relations

   The three variables in this section looked at the family structure of a character and the relative stability of familial relationships.

   Most of the core sample (40%) was characterized as a member of a “traditional family” -- husband, wife and child(ren) -- regardless of which role the coded character happened to be assigned. 15% of core-characters had no known family and 13% were part of a single-parent family. This family structure, though, was unknown for about one-fifth of all characters. The size of this unknown group might be attributable to those plot-functional, but, nonetheless, more minor characters who were not featured in all episodes in the season. With such truncated roles, it is unlikely that there was much narrative need for revealing such characters’ families. Despite the fact that 60% of core characters were “understood” to be part of some sort of family unit, it was found that only 25% of the time were they ever in the presence of their family. Thus, it might be interpreted that, in a sense, family situations often function like the classroom or the workplace: we (the viewers) “know” that these things are part of the characters’ lives, but they are, nonetheless, not that frequently built into the situations actually dramatized.

   When a character was shown to interact with family members, the well-being or cohesiveness of family interaction was coded. The mode (the value that was coded the most) of a character’s family interaction over all episodes was used for the core-character measurement of family interaction.

   Because familial relationships were often not indicated, almost half of core characters could not be coded in this respect. Still, almost one-third of core-characters enjoyed typically smooth family interaction, about 15% of core-characters had mixed interaction, and about 3% were dramatized in the context of more fractious (rocky) families.

   The small amount of family interaction may again be seen as a function of genre: there are “school dramas” or “family programs” or “friendship-based comedies”, but few programs represent all these contexts.

Personal Success

   A character’s personal success was assessed in terms of six variables that are organized around two distinct areas: (1) interpersonal interaction and (2) psychological well-being.

   Three separate variables pertained to interpersonal interaction: likeability, popularity, and quality of friendships. Again, core-character measurements relied on finding the mode condition among the characters’ per-appearances and using that central tendency to characterize the player more generally, For example, if a character was coded as being likeable in 17 of 25 appearances, his/her core-character value was considered likeable. As also noted earlier, it was possible for a character to be plot-functional, but nonetheless be shown in only one or two episodes, thereby not providing adequate characterization so as to determine how other characters “felt about” or interacted with him or her. This would account then, for high percentages in “not clear” or “not applicable” categories.

   Overall, the likeability of about a third of core characters could not be coded. Still, almost half of all core characters were considered likeable, and likeable characters. About 10% of characters were coded as variously liked and not liked (mixed) and fewer than 10% were unlikable.

   Interestingly, perhaps, around 60% of core characters were shown to have no friends. This can again be attributed, most likely, to minor characters seen in relatively few episodes. That friendships don’t seem too prevalent may relate to the fact that only three “friendship-based” shows (i.e. Friends, Beverly Hills 90210, Dawson’s Creek) were included in this sample. Other programs were either “family-based” (i.e., Home Improvement, Party of Five, The Simpsons, and King of the Hill), “work-based” (i.e., ER) or a combination of one or more of these categories -- friendship, family or work (i.e., Sabrina, the Teenaged Witch and Boy Meets World).

   For 54% of core characters either no friendships were represented or what was represented was not sufficiently detailed to code in this context. Keeping this in mind, about 40% of the core characters typically showed smooth friendships (in almost half of all appearances), while fewer than 7% showed uneven (mixed) relationships.

   In addition to interpersonal interaction, the other set of variables pertaining to personal success concerned psychological well-being. This second category included three variables: psychotherapy, self-esteem and overall happiness. The psychotherapy variable was coded in terms of its absence or presence in an appearance - whether or not a character sought or was given counsel at least once during the entire season. With respect to the self-esteem and overall happiness variables, coding was done using a very basic three-point scale: high, average, and low. In both cases, the average value pertained both to characters who were not adequately dramatized as to make such inferences or to characters who were dramatized without any noticeable reference to their, depending on the case, self-esteem or happiness, high or low. The mode of the per-appearance values for these variables was then used for core-character measurements of self-esteem and overall happiness.

   Psychotherapy was sought and/or obtained by six core characters (fewer than 5%) in eight separate appearances.

   Over 80% of core characters were coded as having average self-esteem; the approximately 15% remaining core characters were judged as having high self-esteem, and fewer than 3% were coded as having low self-esteem. Not only were characters generally not shown in the extreme with respect to self-esteem issues, but most characters (over 90%) were coded as being average in terms of overall happiness.

   As all the variables pertaining to interpersonal interaction and psychological well-being (considered together here as personal success) indicate, although certain aspects of characters’ interpersonal lives were often not revealed, when such revelations were made, major characters were typically not shown to be interpersonally or psychologically extreme. One might imagine that characters in teen-related shows would be more angst-filled than these patterns suggest, and, to that extent, the data from this section might be viewed with some curiosity.

   One reason that this examination may not have found more emotionally-charged characters may, in fact, be an artifact of the coding instrument. Typically, the coding instrument was not completed until the end of each sampled episode (even though notes were recorded while viewing). Thus, the “condition” of the character at the end of the episode may have been particularly influential in determining how variables were coded. More specifically, given that 60% of the programs in this sample were comedies -- and thus, invariable end on a positive note (Thomas & Callahan, 1982) -- this may have had some significant influence on this study’s findings with respect to the psychological well-being variables in particular. Perhaps a finer delineation would have been more effective, so as to indicate if, in the course of an episode, a character ever appeared unhappy/upset, doubted him/herself or, conversely, felt extremely happy or excited.

Romance and Sexual Behavior

   The relational status of characters was coded as: single, significant other, cohabitating, engaged, or married. Single characters comprised about half of the core characters. Almost 15% of core characters, had significant others; over 20% of core-characters were married. Together, cohabitating and engaged comprised almost 10% of both the core-characters.

   These shows, then, most commonly portrayed single people. It is arguably a shortcoming of this research that the coding instrument did not distinguish between single characters who were actively dating and single characters who were not. This distinction may be particularly salient with respect to adolescent and teenaged characters as well as viewers, neither of whom are likely to be engaged or married. This saliency relates more specifically to the putatively common assumption that television viewing is an activity much engaged in by the dateless; thus, the representation of similarly dateless characters both in terms of frequencies and characterization, as well as the comparative portrayal of dating versus non-dating characters, might be of some special interest when particularly concerned with adolescent and teenaged viewers.

   Another reason to distinguish between dating and non-dating characters would relate to the issue of sexually-transmitted diseases. This study, though, does in some respects attend to this issue by studying the number of sexual partners attached to each character.

   Unless a character was coded as single, the relative stability of his/her relationship was broadly distinguished as either stable or rocky. There were 58 of the 144 core characters who were single, or whose relational status was unknown. Of the core characters with romantic partnerships, over two-thirds were shown as having stable relationships.

   Clearly there are two contexts in which one would interpret these data as “positive” or “optimistic” with respect to portraying romance. The simpler of the two is that the vast majority of relationships (two-thirds) are good ones, i.e. stable. The somewhat more complex analysis pertains to what Thomas (1994) calls the “political-realism” perspective, i.e. comparing media statistics to those culled from real-life. In this context, a 70% relationship-success rate is essentially “over-representative” insofar as in real life nowhere near that many marriages are successful. Moreover, assuming that the success rates for marriage would be greater than those for all forms of non-institutionalized relations (from dating to engagements), the 70% success rate on these shows, then vastly “over-represents” real-life probabilities. Of course, on the one hand, since this study only addresses one season of these shows, it may be that certain relationships that were successful during this season, eventually failed in subsequent seasons. On the other hand, it might also be argued that since television drama is not performed in classical time, it is impossible to tell the longevity of these partnerships in the context of their dramatic world.

   The sexual history variable assesses characters “sexual-partnering” behavior, as represented over the course of the television season. That is, a consolidated or cumulative index was used to indicate the furthest level of sexual intimacy in which a character engaged over the season, as well as the number of partners with whom s/he engaged in some form of sexual intimacy. The five possible designations for a character’s sexual history are indicated in Table 1 which also presents their frequencies. [INSERT TABLE 1]

   As shown in Table1, over 40% of all characters in the sample never engaged in any sexual activity. Approximately one-quarter of all characters engaged in only kissing and necking during the season, with the vast majority doing so with but one partner. Another one-fifth of the characters did engage in sex during the course of the season, but with only one partner. Few characters (8.3%) had sex with more than one partner. In this sample, at the very least, on shows most popular among adolescents and teenagers, characters cannot aptly be described as unduly or intensely sexual; in terms of their sexual histories, approximately two-thirds of all characters never engaged in sexual activity more intense than kissing and necking, and indeed, of this two-thirds, most don’t engage in any sex whatsoever.

CONCLUSION

   The present study set out to answer the research question: “Who inhabits the world of adolescent television and what are the norms and values of its “culture”?