Final episodes of long-running and greatly loved television series achieve famously high ratings (Battaglio, 2001). It was hardly surprising, then, that an estimated 51 million viewers tuned in to view the final episode of Friends, which aired in the United States on May 6, 2004 (Associated Press, 2004). Although viewers were no doubt aware that they would be able to see their friends from Friends over and over again in reruns and DVDs, the last episode seemed to mark a farewell of some import to many millions. The vast majority of viewers know that their relationships with television characters are imaginary (Caughey, 1985), and yet, as the ratings numbers and the general commotion around this and other finale shows suggest, the end of such relationships is emotionally meaningful. What do viewers feel when relationships with television characters come to an end? To what extent are separations from television characters similar to endings of personal relationships? What factors impact the intensity of feelings associated with such breakups? Which viewers experience these feelings more strongly than others? This study attempts to answer these questions with data collected from viewers immediately after the end of Friends.
This study is set within the framework of parasocial relationships (PSRs). Initially defined by Horton and Wohl (1956) as a "seeming face-to-face relationship between spectator and performer" (p. 215), PSRs have been widely studied, both in terms of their development and in terms of their influences on viewers' emotional states and reactions to television exposure (e.g., Auter, 1992; Eyal 8, Rubin, 2003; Turner, 1993). PSRs are now understood to be an integral and important part of many people's systems of social relationships and "the distinction between social and parasocial relationships, which Horton and Wohl  assumed was so obvious, is increasingly complex and hard to define" (Cohen, 2004, p. 200). As discussed later, the topic of PSRs is, in fact, now recognized as a potential contact point between mass media and interpersonal theories (Turner, 1993). Researchers are increasingly applying interpersonal, relational, and developmental theories to the study of PSRs (Cohen, 2003; Cole & Leets, 1999; Isotalus, 1995). This study contributes to this literature by applying aspects from theories of relational development to the study of people's parasocial relationships with mediated characters. It extends this literature by examining the application of theoretical premises regarding relational dissolution to the study of the termination of imaginary relationships.
Friends came on the air in 1994 following NBC's success with Seinfeld, and like its predecessor, was created as a sitcom set not in a family home or business, but rather focused on a group of young single adults. In an age of segmented viewing when the viewing unit is no longer composed solely of nuclear families, the time was ripe to experiment with moving the focus of sitcoms away from families. Furthermore, a program about young, urban singles made sense based on the belief that viewers relate and identify with those who are similar to them and the special attractiveness of the 18-to-30 demographic to advertisers. However, unlike Seinfeld, famous for being a show about "nothing" (CNN, 1998; TV Tome, 2005), Friends was a show about something: It explored the interpersonal relationships of its stars as a basis for its plot and humor. This heightened the potential for viewers to feel like they were a part of this group of friends, a feeling Auter and Palmgreen (2000) showed to be an important part of relationships with the characters. Over 10 years viewers were invited to watch these six friends interact, learn about them in intimate and meaningful ways, and vicariously experience the trials and tribulations of young adulthood. Most of the college students who took part in this study were still in elementary school when the show first aired and grew up watching the show. It is thus not surprising that the show's ending would be an emotional experience for many of them.
As the significance of PSRs in the process of media influence has become more apparent (Basil, 1996; Brown, Basil, & Bocarnea, 2003; Papa et al., 2000), researchers have become interested in exploring such relationships and understanding how they impact viewers. Somewhat to their surprise, researchers have consistently found that relationships with television characters do not replace relationships with friends, but rather complement social relationships (Kanazawa, 2002; Perse & Rubin, 1990; Tsao, 1996). Feelings toward television characters do not generally serve as a replacement for primary social relationships but rather keep one company (Isotalus, 1995) and like ordinary friendships serve to provide people with social enjoyment and learning.
PSRs are a set of feelings viewers develop toward media characters that allow viewers to think and feel toward characters as if they know and have a special connection with them. These feelings extend beyond the moment of viewing (Horton & Wohl, 1956) and continue from one viewing situation to the next. Such relationships originate from repeated viewing of characters that simulate social interaction, and they develop and strengthen over time (Isotalus, 1995; Perse & Rubin, 1989; R. B. Rubin & McHugh, 1987; but see also Auter & Palmgreen, 2000). As viewers are exposed to characters over a longer period of time and more frequently, they develop more confidence in their attribution of how the character will behave and have less uncertainty in their relationships.
Television characters provide viewers with one-way relationships, and the intimacy they offer is, as Horton and Wohl (1956) argued, only at a distance. Nonetheless, Koenig and Lessan (1985) found that viewers rated favorite television characters as further from themselves than friends but closer than acquaintances. Newton and Buck (1985) concluded their findings by suggesting that television can be seen as a significant other. Thus, television personalities are a significant part of one's social network, although their social and emotional functions seem to be limited compared to close family and friends.
In terms of their effects, Fisherkeller (1997) suggested that at least for some teens, media characters serve as models for how to achieve goals that are related to the development of their identities. Other scholars have shown that imaginary relationships with media characters have real social consequences, such as increasing the persuasive power of public service announcements when they feature celebrities with whom viewers have PSRs (Basil, 1996; Brown et al., 2003). Similarly, Sood and Rogers (2000) linked the effects of education-entertainment programming to the development of PSRs with soap opera characters. Most recently, Schiappa, Gregg, and Hewes (2005) found that parasocial contact and relationships can change attitudes about homosexuality. The importance of mediated relationships, and their similarity to social relationships, therefore, suggests that the end of a long-standing and popular show like Friends should also be significant and that it may be a cause of some distress. It also remains to be seen whether and how the one-way and distant nature of such PSRs leads to differences in the responses to their end.
The notion of parasocial breakup (PSB; Cohen, 2003) describes a situation where a character with whom a viewer has developed a PSR goes off the air. This may happen because a show ends, because a character is taken off the show, or because something happens to the actor or actress who plays the character. In turn, a viewer may decide to stop watching the show or become less interested in or less devoted to the character.
The dissolution of close social relationships has been found to lead to depression and is a common reason for seeking psychological counseling (McCarthy, Lambert, & Brack, 1997). In regard to celebrities, Meyrowitz (1994) described extreme reactions exhibited at the death of celebrities such as Elvis Presley and John Lennon. Based on his analysis of these extreme cases and his discussion of more general patterns of responses to the death of what he called "media friends," he concluded that, "these relationships have features that are very human, very warm, and very caring" (p. 80). Although the myths, rituals, and pilgrimages that have come to surround the death of media megastars do not characterize common responses to the end of most television series, they do point to the emotional potential of imaginary relationships.
Research has found that though the dissolution of parasocial relationships is less stressful than that of close relationships, it follows some similar patterns (Cohen, 2003, 2004). Cohen asked respondents to imagine how they would feel if their favorite television persona would be taken off the air. He found that like in social relationships the stress of (imagined) breakup was strongly related to the intensity of the relationships. However, women, who generally report stronger PSRs (e.g., Tsao, 1996), did not report expecting higher levels of distress if their favorite television personality went off the air (Cohen, 2003). This finding echoes the fact that, although women tend to have stronger interpersonal relationships, they are better able to cope with the end of these relationships (Helgeson, 1994; Simpson, 1987; Sprecher, Felmlee, Metts, Fehr, & Vanni, 1998). Teenagers who are generally seen as more emotional and more involved with media characters than adults were also found to expect stronger emotional responses to PSB than adults (Cohen, 2003).
These studies were able to confirm the idea of PSB by showing that people expected to be sorry when their favorite character went off the air, and to establish a basis for comparing PSB to social...