Although interest in the psychology of the sport spectator dates back to the early 1900s (Howard, 1912; Patrick, 1903), empirical research on this population is relatively rare. In fact, a review of research published in sport psychology and sociology journals between 1987 and 1991 revealed that only 4% of the studies examined spectators (Wann & Hamlet, 1995). The research and writing which has been conducted has usually focused on either spectator aggression (e.g., Branscombe & Wann, 1992a, Russell, 1993) or mood (e.g., Hirt, Zillmann, Erickson, & Kennedy, 1992, Sloan, 1979, Wann & Branscombe, 1992).
A significant void in the research lies in the area of seasonal changes in spectators' identification with their favorite team. That is, no studies employing a longitudinal design had been conducted, although researchers had acknowledged a need for such work. In fact, Murrell and Dietz (1992) stated that "One important direction for future work is to determine whether the impact of fan identification is indeed resistant to the, outcome of social competition" and further, that the method of choice for examining this topic would be "assessing the changes in that identification over time and with wins and losses produced by the team" (p. 36). This design would allow for the examination of within-subject changes in response to the team's successes and failures and, as such, would shed light on the psychology of the sport spectator. Such a research endeavor was the focus of this work.
Research investigating the identification of spectators is important for a number of reasons. First, because degree of fan identification is a major predictor of fan violence (Branscombe & Wann, 1992a, Simons & Taylor, 1992) and evaluations of other spectators (Wann & Dolan, 1994a), an understanding of seasonal changes in fan identification may assist in the understanding and control of the violence. Similarly, because identification seems to be related to other arousal and affective reactions (Branscome & Wann, 1992b; Hirt et al., 1992; Wann & Branscombe, 1992), research illuminating the processes underlying fan allegiance should shed light on how fans react emotionally to the exploits of their team. And finally, because fan identification has also been found to moderate other fan behaviors including tendencies to distance oneself from the team (Wann & Branscombe, 1990), attributional processes (Wann & Dolan, 1994b), and self-esteem (Branscombe & Wann, 1991), information concerning changes in team identification should further the understanding of a variety of fan behaviors.
It appeared that several variables may play a role in possible changes in identification. The first such variable was identification level itself, that is, the extent to which the spectator feels allegiance to the team. Because research endeavors such as those described above found that level of identification is a major predictor of a variety of spectator behaviors, it appeared possible that spectators with differing levels of identification with the team may exhibit a differential pattern of responses throughout a season. For example, highly identified persons may be rather consistent in their allegiance while low identified persons exhibit increasingly lower levels of allegiance as the season progresses. This possibility was examined in the current research.
In addition to identification level, two other factors may be important in altering the identification of spectators: the outcome and the location of the team's previous game. As for competition outcome, this variable has been found to be vital in determining the postgame emotional reactions of fans (SIoan, 1979, Wann, Dolan, McGeorge, & Allison, 1994). Although post-game differences in identification as a function of outcome were not found in one study (Wann et al., 1994), this research was conducted using a between-subjects design and …