Sociology of Sport Online School of Physical Education, University of Otago

 [Home | Contents | Forthcoming | Overview | Submit]

Geographical relocation, suicide, and homicide: An exploratory analysis of the geographic relocation of professional sports teams in three U.S. areas and the impact on suicide and homicide rates

Robert M. Fernquist
Central Missouri State University
Department of Sociology and Social Work
Wood Hall 203
Warrensburg, MO 64093
June 20, 2001
Phone: (660) 543-8510 fax: (660) 543-8215

Key Words: suicide, sports, geographical relocation


Tutko (1989) has argued that the performance of sports teams can greatly impact fans’ feelings about themselves and about life. Wann, Melnick, Russell, and Pease (2001) have found that fans who identify strongly with their team have difficulty psychologically distancing themselves from ‘their team’ when the team does not perform well. While the performance of professional sports teams in relation to personal violence (e.g., suicide and homicide in the general population) has begun to be studied, I examine a new angle on the relationship between professional sports teams and personal violence-that of geographical relocation. In this exploratory analysis, I look at the geographical relocation of professional sports teams from three areas: New York in 1957 (the Brooklyn Dodgers and New York Giants’ moves), Cleveland in 1995-1996 (the Cleveland Browns’ move), and Houston in 1997-1998 (the Houston Oilers’ move). Results indicate that, while homicide rates seem to be unaffected by these moves, suicide rates appear to increase.


It is well known that sports fans often ardently track the successes and failures of ‘their’ team(s). Tutko explains that To see a performance where the (team) does well is tied to the ‘good’ feeling inside of us-that feeling of trying and doing well. We assume that the good feeling is tied to all of the positive characteristics that make up character (1989, p.115). Further, sports fans often reflect on their own trials and pressures when watching professional sports because

"athletics . . . excites our dominant personality features" (Tutko, 1989, p.115).

Branscombe and Wann (1992a, p. 1017) write that, for fans who identify strongly with ‘their team’, “An affront or loss on the part of the team then is a loss for the self.” Building on this concept, Mitrano (1999, p. 151) explains that, just as people tend to go through certain stages (such as anger and depression) when dealing with death or some other significant loss, “. . . fans collectively experience the same stages of grieving and loss” when a pro sports team relocates to another geographical area.

Fernquist (2000) has found that the performance of professional sports teams is related to local suicide and homicide rates, whereas strikes in professional sports are related to national homicide rates (2001). In this exploratory research, I examine how, if at all, the relocation of professional sports teams from one geographical area to another impacts persons in the area left behind. Specifically, I examine how the relocation of the Brooklyn Dodgers and New York Giants after the 1957 baseball season, of the Cleveland Browns after the 1995-1996 football season, and of the Houston Oilers after the 1997-1998 football season impacted suicide and homicide rates of persons in New York City, Cuyahoga county, and Houston city, respectively. If there is much merit to the idea that some sports fans become closely and emotionally attached to ‘their team’, and there is much research to support such attachments (as will be discussed below), then the examination of suicide and homicide rates among the general population in the areas left behind warrants investigation. I use Gabennesch’s (1988) theory of broken promises and Durkheim’s concept of anomie (1897[1951]) as theoretical guidelines for our research.

Literature review and theory

In the following literature and theoretical review, I discuss (1) sports fans’ identification with sports teams; (2) the impact that franchise relocation, as well as sport teams in general, have on society; and (3) sociological theories related to suicide, homicide, and the world of sports. Sports fans’ identification with sports teams: Although the intensity of fan identification with sports teams varies from fan to fan (Wann, 1996), research still finds that following sports teams has significant impacts on the lives of the fans. In this respect, Pooley (1980, p. 20) stated that “being a fan usually means feeling intensely about a team as well as following its activities”, while Cialdini, Borden, Thorne, Walker, Freeman, and Sloan (1976) similarly reported that college students tend to bask in the reflected glory (BIRG) of their respective football team's victories. Hence, "through their simple connections with sports teams, the personal images of fans are at stake when their teams take the field. The teams’ victories and defeats are reacted to as personal successes and failures" (Cialdini et al., 1976, p. 374). Branscombe and Wann (1991) found strong evidence to suggest that many sports fans like following ‘their team’ because this allegiance provides them with pleasurable leisure activities and increases their self-esteem and overall positive emotions. If such positive outcomes can result from being a sports fan, why, then, could identification with a sports team result in aggressive and/or suicidal behavior? 

In answer to such a query, Branscombe and Wann (1991, p. 125) have argued that the degree of threat to a social identity, combined with an inability or unwillingness to reduce the importance of that identity, . . . is what gives rise to the negative potential involved in sports team identification. When the threat is at reasonably low levels (for example, when a championship is not at stake), then only the positive effects of attachment to something larger than the self will be observed. Both sides of the coin, however, can have powerful social consequences. Branscombe and Wann (1992b) also reported similar results regarding fan identification and sports. Further, it is common for “fans with a high degree of team identification” to be unable to distance themselves from their team’s failures because “the concept of team follower is such a central component of (their) self-identity (Wann, Melnick, Russell & Pease 2001, p. 171).

These fans keep hoping and waiting for their team to do better and become frustrated when their team continues to perform poorly. There exists, therefore, sports fans who become so enmeshed with their team that these fans’ “psychological well-being . . . is jeopardized” when their team is not doing well (Wann et al. 2001, p. 172). Branscombe and Wann (1992a, p. 1017) further explain that when sports fans cannot or will not distance themselves psychologically or emotionally from a sports team, “. . . aggressive behavior may be resorted to as a means of identity restoration”. Such aggressive behavior may surface in the form of suicide or homicide.

Based on the review of these studies, the benefits derived from identifying with their team (i.e., increased positive emotions, higher self-esteem, enjoyable leisure activities) can, for fans who strongly identify with their team, result in problems. These problems (such as aggressive behavior and/or depression) arise because team identification is such a central a part of their sense of self that when ‘their team’ fails they find themselves unwilling or unable to distance themselves from the team in hope, perhaps, that the benefits will quickly return to them.

The impact that franchise relocation and sport teams in general have on society

Mitrano (1999) examined fan reaction to the franchise relocation of the Hartford Whalers hockey team during the period from March of 1996 to May of 1997. The last season the Whalers played in Hartford was the 1995-1996 season. Mitrano likened the grief and loss expressed by many of the fans he interviewed over the franchise relocation to grief and loss people commonly go through when a loved one dies or some other significant loss is encountered. Some of the comments fans offered ranged from “It’s really over. I’m truly too numb to really say much” to “Now that  our team is leaving, ask yourself, ‘Will you stay in CT?’ I’m not” to “I’m completely numb . . . . my life has changed” (Mitrano,1999, pp. 138-140). Although not all sports fans reacted to the franchise relocation in this manner, Mitrano (1999) indicated that a sense of loss and grief were common reactions among the fans he studied. Another common occurrence noted in this study was “. . . a shift from the tenets of loyalty and stability” among fans toward both team management and sports teams in general to tenets of “infidelity and celerity” (Mitrano, 1999, p. 151). Franchise relocation, then, cannot only make fans change their feelings toward their team, but can also negatively impact their sense of self. 

Regarding the franchise relocations involved in the present study, Sullivan (1987 p. vii) explains that the Brooklyn Dodgers “ . . . enjoyed a close emotional tie with” the people in Brooklyn. Sullivan further explained that the Brooklyn Dodgers had such a “. . . powerful emotional” effect on Brooklynites that one New York Times writer expressed the mood of the Brooklyn fans as ‘galling resentment’ when they learned their team would be leaving Brooklyn and going to Los Angeles after the 1957 season (Sullivan, 1987, p. 137). The Browns moved from Cleveland following the 1995-1996 football season; a move that was similarly upsetting to fans. For example, fans picketed Cleveland Stadium as owner Art Modell was expected to announce the move from Cleveland (USA Today, 1995). However, unlike the widespread passion that existed when the Browns and Dodgers moved, there was more ambivalence on the part of fans when the Oilers moved away from Houston after the 1997-1998 season (Collegian 1996). Clearly, though, there were fans in each of these areas that were upset over the geographical relocations of their pro sports teams.

Ecological studies suggest that fans are impacted by the performance of professional sports teams. Trovato (1998) found that when the Montreal Canadiens hockey team was eliminated in an early round from the playoffs between 1951 and 1992, the likelihood of suicide increased for males aged 15-34 in Quebec. Similarly, Fernquist (2000) reported that suicide rates in metropolitan areas tend to be lower when those metropolitan areas have professional sports teams that performed well (e.g., made the playoffs regularly and won championships). He further found that the more times a team in a given metropolitan area made the playoffs, the lower the homicide rate in the given metropolitan area. Fernquist (2001) also argued that, during the baseball and hockey strikes of 1994-1995, national homicide rates increased while suicide rates were not significantly affected. Both Trovato’s and Fernquist’s work suggests that social integration among a certain segment of the population (i.e., fans who strongly identify with ‘their team’) is significantly impacted by the performance of professional sports teams. The impact that the geographical relocation has on personal violence among persons ‘left behind’ is as of yet unknown. Given the manner in which suicide and homicide are related to phenomena in professional sports, as described above, the study of suicide and homicide in relation to the relocation of professional sports teams is worthy of further investigation.

Sociological theories related to suicide, homicide, and the world of sports: Gabennesch (1988) developed a theory in which he describes how disappointments may result from expectations not being met - disappointment which could result in suicide. Gabennesch argued that suicide is likely to occur when promises are broken due to "conditions or events which induce psychological misery" (1988 p.142). Gabennesch’s theory of broken promises does apply to professional sports since persons who expect ‘their’ team to do well may become extremely disappointed when the ‘promise’ of winning is broken (Fernquist 2000). Stack (1995) argues that Gabennesch's theory of broken promises can also be applied to homicide, since disappointment was central to Durkheim’s (1897[1951]) short discussion of anomic homicide. Durkheim (1897[1951], p.357) argued that anomie promotes "a state of exasperation and irritated weariness" which could result in either suicide or homicide, depending upon the person's "moral constitution" (e.g., a person with low morality is more likely to commit homicide than suicide).

Although some sports fans tend to have a long-term ‘love affair’ with their sports teams (Wann 1996), I use Durkheim’s concept of anomie in addition to Gabennesch’s theory of broken promises to guide this research. Durkheim (1897[1951]) explained that sudden (or anomic) changes in the environment can have an adverse impact on people. Durkheim, in his classic sociological analysis of suicide, showed that sudden changes such as widowhood, divorce, and economic crises can drastically change the way people view the normative order in society. When such sudden changes occur, it takes time for the person to adapt himself “to the new situation in which he finds himself and accordingly offers less resistance to suicide” (p. 259). Anomic suicide, therefore, “results from man’s activities lacking regulation and his consequent sufferings” (p. 258). Within this context, I acknowledge the research done by Curtis, Loy, and Karnilowicz (1986) in which they examined social integration in relation to professional sporting events. Curtis et al. (1986) found significant, albeit weak, declines in suicide just before and during the last day of the World Series and Super Bowl Sunday compared to the days immediately following these sporting events from 1972-1978. Smith (1976) also reported that interest in professional sports teams/persons can increase fans’ social integration1. However, I test Durkheim’s theory that sudden changes to one’s environment can have detrimental effects. The sudden changes central to this study are related to the geographical relocations of the sports teams. Even though fans generally know the geographic relocation is imminent, the anomic nature of the end of the last season is what causes problems since fans feel they no longer have ‘their team’ in ‘their area’ to support. I argue that the end of the final season, not the actual move itself, is what is anomic since the end of the season for all intents and purposes is what signals to fans that their team has ‘gone’. I further argue that the beginning of the next season (the first season following the move) is anomic to fans since fans are once again reminded that ‘their team’ is no longer in their geographical area. I hypothesize, therefore, three contexts which increase suicide and homicide rates among persons in the area ‘left behind’: (1) increases that occur at the end of the last season before the actual move occurs; (2) increases that occur at the beginning of the next season; and (3) increases that occur at both the end of the last season and the beginning of the next season. I will first examine descriptive statistics regarding the frequency of the number of suicides and homicides in relation to these hypotheses and then, based on these trends, I will estimate regression models to predict variation on suicide and homicide rates. Although such an approach to data analysis can be termed ad hoc, an examination of trends before the statistical modelling will allow us to see if trends present in the descriptive analysis remain in the presence of relevant statistical controls.

Data and Methods

For each geographical area, I define the end of the season as occurring during December and January for American football and September and October for baseball (when the regular season ends and playoffs occur). I define the beginning of the next season as occurring during August and September for American football and March and April for baseball (when pre-season games come to an end and the regular season begins). Monthly data for all areas are used in this analysis. To control differences in the seasonality of suicide, I compare data for the months at the end of the season and beginning of the next season (e.g., the ‘anomic months’) with those same months in the two years before and the two years after the ‘anomic months’. The central variable of interest in this analysis, the ‘relocation’ variable, is used in regression analysis and is a dichotomous variable, being coded 1 for anomic months and 0 for all other months. The coding of this variable is explained in more detail below. The Brooklyn Dodgers and New York Giants’ moves: Data on the number of suicides and homicides in New York City are from the New York City Department of Health (2000). Data for individual boroughs were unavailable. For the last season’s end, the anomic months are September and October 1957 and I compare these months to September and October, 1956, 1958, and 1959 (1955 data were not available). For the beginning of the season after the baseball teams moved, the anomic months are March and April, 1958, and I compare these months to March and April, 1956, 1957, 1959, and 1960.

The Cleveland Browns’ move: Data in this instance were obtained from Cuyahoga county, the county in which the city of Cleveland is located. Data on suicides and homicides in Cuyahoga county are from the Ohio Department of Health, Center for Public Health Data and Statistics (2000). For the last season’s end, the anomic months are December 1995 and January 1996, and I compare these months to December 1993, 1994, 1996, 1997 and to January 1994, 1995, 1997, and 1998. For the beginning of the season after the Browns moved, the anomic months are August and September 1996, and I compare these months to August and September 1994, 1995, 1997, and 1998. 

The Houston Oilers’move: Data on the number of suicides and homicides in Houston city are from the Texas Department of Health (2000). For the last season’s end, the anomic months are December 1997 and January 1998, and I compare these months to December 1995, 1996, 1998, 1999 and to January 1996, 1997, and 1999 (2000 data were not available). For the beginning of the season after the Oilers moved, the anomic months are August and September, 1998, and I compare these months to August and September 1996, 1997, and 1999. 

In the regression models, I examine how the geographic relocations impact suicide and homicide rates per 100,000 population. The central independent variable in the regression models, the ‘relocation’ variable, is a dichotomous variable comparing suicide and homicide rates in the anomic months to the same months in the years preceding and succeeding the anomic period. To convert the number of suicides and homicides to rates per 100,000 population for the regression analysis, I obtained population estimates for the three areas. In the regression models, birth rates, marriage rates, and unemployment rates per 1,000 population are controlled for, variables which are relevant to the sociological study of suicide and homicide (Fernquist 2000; Lester 1994). For New York City, data on the number of births and marriages, as well as estimates of population size, are from the New York State Department of Health (2000). Data on unemployment  rates for New York City are from the New York State Department of Labor (2001). Unfortunately, unemployment data prior to 1958 were not available, and so 1958 data are also used for 1956 and 1957. For Cuyahoga county, data on the number of births and marriages are from the Ohio Department of Health, Center for Public Health Data and Statistics (2000),while the unemployment rates are from Labor Market Information (2000). The population estimates are from the Ohio Department of Development (2000).

For Houston, data on the number of births and marriages were available only for Harris county (the county in which the city of Houston is located), not for the city of Houston proper. Data on the number of births and marriages are from the Texas Department of Health (2000). Unemployment rates for the city of Houston are from Economagic (2001). Population estimates for persons in Harris county (to calculate birth and marriage rates) are from the Texas Department of Health (2000) while population estimates for city of Houston (to calculate suicide and homicide rates) are from the Texas State Data Center (2000).

Ordinary least squares estimation (OLS) is used to model variation in suicide and homicide rates. Given that the OLS regression models in table 2 employ cross-sectional, time-series data, I checked for harmful autocorrelation (significant correlations among the disturbances) and heteroskedasticity (unequal variances among the disturbances) since these problems may produce incorrect OLS regression coefficients (Stack 1992).

I also checked for problematic collinearity among the predictors. After the suicide and homicide models were corrected for first-order autocorrelation, the regression models were acceptable in terms of autocorrelation and heteroskedasticity levels (see table 2). Since the value of the Durbin-Watson statistic, a commonly used test to examine the degree of autocorrelation present in regression, is close to 2.0 in both models, and since neither statistic that measures the presence of heteroskedasticity (the Glejser [1969] test) is significant at the .05 level, we can be confident that neither autocorrelation nor heteroskedasticity will bias results from the regression models. Allison (1999) explains that the use of Variance Inflation Factors (VIFs) can show whether or not collinearity among the independent variables will bias regression estimates. Each predictor in the models in table 2 was regressed on all others simultaneously to calculate the VIF for each predictor. Each VIF is calculated by 1/1-R2 and Allison (1999) explains that any VIF of 2.5 or larger indicates problematic collinearity. No VIF was 2.5 or larger, suggesting that collinearity was not a problem for the models in table 2.


Table 1 shows the number of suicides and homicides and the percentage changes in the anomic months versus those same months before and after the anomic months. Based on these percentage changes, I will categorize New York City and Houston as being anomic at the beginning of the next season since percentage changes in suicide and homicide are greater at the beginning of the next season relative to the end of the last season. I put Cuyahoga county in the category of being anomic at season’s end since, especially for suicide, percentage changes were greater at season’s end relative to the beginning of the next season. I feel this categorization best explains increases in suicide and, to a lesser extent, homicide in each of these areas.

Table 1 Suicide and Homicide in Three U.S. Areas: Raw Numbers and Percentage Change Over Previous Period

  New York City     Cuyahoga County     Houston City  

Comparisons at End of Season (Annual Averages)

Before Anomic After Before Anomic After Before Anomic After
Sept/Oct 1956 Sept/Oct 1957 Sept/Oct 1958-9 Dec 1993 - 4 Jan 1994 - 5 Dec 1995 Jan 1996 Dec 1996 - 7 Jan 1997 - 8 Dec 1995 - 6 Jan 1996 - 7 Dec 1997 Jan 1998 Dec 1998 - 9 Jan 1999
Suicides 87 96 (+10%) 116 (+21%) 29 33(+14%) 20 (-39%) 43 49 (+14%) 41 (-16%)
Homicides 40 46 (+15%) 65 (+41%) 25 20 (-20%) 21 (+5%) 61 48 (-21%) 48

Comparisons at Beginning of Next Season (Annual Averages)

Before Anomic After Before Anomic After Before Anomic After
Mar/Apr 1956-7 Mar/Apr 1958 Mar/Apr 1959-60 Aug/Sept 1994-5 Aug/Sept 1996 Aug/Sept 1997-8 Aug/Sept 1996-7 Aug/Sept 1998 Aug/Sept 1999
Suicides 102 148 (+45%) 136 (-8%) 27 20 (-35%) 34(+70%) 43 52 (+21%) 35 (-33%)
Homicides 44 60 (+36%) 65 (+8%) 36 25 (-31%) 20 (-20%) 53 61 (+15%) 42 (-31%)


In the regression analysis, I examine suicide and homicide rates per 100,000 population in the three areas combined with the relocation variable coded 1 in New York City and Houston for the anomic months during which the respective sports seasons began the season after the geographic relocations while in Cuyahoga county the move variable is coded 1 for the anomic months at the end of the last season before the geographic relocation. All other months are coded 0.

The regression models in table 2 show that the geographic relocation of all three teams is significantly related to increased suicide rates in New York City, Cuyahoga County, and Houston City, holding all else constant in the model. In Cuyahoga County, the impact on suicide was more immediate (since Cuyahoga was modeled as a ‘season’s end anomic’ area), while the impact on suicide in New York and Houston was delayed until the next season began. Geographic relocation was not significantly related to variation in local homicide rates. Regarding the control variables, birth rates are positively related to both suicide and homicide rates, while marriage rates are inversely related to both suicide and homicide rates. Unemployment rates are not significantly related to either suicide or homicide rates, even though the coefficient in the homicide model is larger than the other coefficients. Explained variance is fairly high in both the suicide and homicide models (.53 and .74, respectively).

Table 2: Unstandardized Regression Estimates of Suicide and Homicide Rates in Three U.S. Areas (N=30)

Independent Variable Suicide Rates Homicide Rates
Move 3.23** 1.07
Birth Rates 0.34** 1.02***
Marriage Rates -0.27* -0.97* * *
Unemployment Rates 0.12 1.59
Constant 6.14 -9.62
Method of Estimation Corrected for First-Order Corrected for First-Order
Autocorrelation (Cochrane/Orcutt) Autocorrelation (Cochrane/Orcutt)
Durbin-Watson Statistic 1.98 1.94
Glejser test for 4.49 (p>.34) 7.3 9 (p>. 11)
Heteroskedasticity (df=4)
Largest Variance Inflation Factor 1.41 1.41
Explained Variance 0.53 0.74
p<.05; **p<.Ol; ***P<.001


Discussion and Conclusion

As Trovato has so correctly stated, “The association of major sport occasions and mortality has received scant attention in the literature, and our understanding of this phenomenon is far from complete” (1998, p. 119). In response, within this study, I have attempted to shed more light on the relationship between sporting ‘events’ and suicide and homicide. This exploratory analysis reveals support for the hypotheses that suicide rates increase due to geographic relocation of professional sporting teams both at the end of the last season before the actual move occurred (Cleveland) and at the beginning of the next season after the relocation (New York and Houston).

The hypotheses about homicide rates, however, were not supported. Although these data are not individual-level thus not allowing us to ascertain if those fans who highly identify with ‘their’ team are the persons whose suicide rate increases upon franchise relocation, the results are suggestive that this may be the case. Whereas Durkheim (1897 [1951]) argued that anomie promotes a state of exasperation which could result in either suicide or homicide, I find that exasperation related to geographic relocation is related to increases in suicide rates, but not homicide rates. The sudden change brought about due to the geographic relocations of pro sports teams does appear to, at least for a short time, make highly identified fans drastically change the way they view the normative order in society. Unable (or unwilling) to adapt to the new environment, an environment which no longer contains ‘their team’, their sense of self may become seriously damaged with suicide more likely to result.

They may come to feel as if they are losing control of the world around them (Grossberg 1992). As Tutko (1989) said, fans put much feeling into following the progress of ‘their team’, thus when that team is no longer in the specific area anomie increases. These fans end up having a difficult time letting go of their team because they “experience depression and an intense negative affective state and adopt a poor outlook on life 
subsequent to” the team’s geographic relocation (Wann et al. 2001, p. 172). 

From the point of view of Gabennesch’s (1988) theory, highly identified fans may come to see the geographic relocation of their team as breaking the promise that, at least in the fans’ minds, their team will always be in that geographic area for the fans to cheer and follow. Although there is certainly never a guarantee that a professional sports team will remain in a given area indefinitely, fans apparently hope/expect their team to stay. Although aggregate data cannot be used to make specific individual references, the data from this study, in agreement with Gabennesch (1988), suggest that when the team does move, fans may feel that these expectations are broken, and these broken promises have detrimental effects. 

I also realize that the nature of being a sports fan has changed over the past several decades. With the introduction of free agency in the 1970s, fans have come to realize that players will come and go from team to team (Wann et al. 2001). Since this study entails data from pre-free agency years (the Dodgers and Giants) as well as data after the introduction of free agency, could the nature of fans’ identification with ‘their team’ have changed over the years? I argue that the nature of fans’ identification with their team has not significantly changed during this period since all three geographic relocations were related to increased suicide rates. If fans have become less attached to ‘their’ team, we would not have seen suicide rates rise after the Oilers and Browns relocated. Therefore, even though fans know players will come and go, they still today expect ‘their team’ to remain in its geographical location. I realize that there are other factors related to suicide and homicide rates that were not included in the study. Variables such as religiosity, per capita income, and cohort size are relevant control variables I was unable to include in the analysis due to a lack of data. It is unclear if the inclusion of these control variables in the regression models would alter the impact of geographic relocation on suicide or homicide.

Urban revitalization has also been occurring in the Cleveland metropolitan area during the 1990s and this phenomenon could also be responsible for significant variation in the Cuyahoga county suicide and homicide rates. Therefore, I acknowledge shortcomings in the current aggregate analysis on suicide and homicide rates in the New York, Cleveland, and Houston areas. Nonetheless, this research does suggest that the association between geographic relocation and suicide (and even homicide) should be studied in more detail. In particular, due to the difficulty highly identified fans have of letting go of ‘their team’ (Wann et al. 2001), it is recommended that the emphasis of future research on geographic relocation and suicide should be placed on the differences in suicidal behavior between highly identified fans and less identified fans.

Thanks are expressed to the following persons for the generous help in providing the author with vital data: Donna Smith (Cuyahoga county); Anna Vincent (Houston city and Harris county); Elizabeth Acevedo, Louise Berensen, Robert Roddy, and Al Wolf (New York City). Thanks also go to Joyce Knapton and Charlene Zion for further help with obtaining data and to the anonymous reviewers of sosol for insightful comments.


Allison, P.D. (1999). Multiple regression: A primer. Thousand Oaks, CA: Pine Forge Press.

Branscombe, N.R & Wann, D.L. (1991). The positive and self-concept consequences of sports team identification. Journal of Sport and Social Issues, 15, 115-127.

Branscombe, N.R & Wann, D.L. (1992a). Role of identification with a group, arousal, categorization processes, and self-esteem in sports spectator aggression. Human Relations, 45, 1013-1033.

Branscombe, N.R & Wann, D.L. (1992b). Physiological arousal and reactions to outgroup members during competitions that implicate an important social identity. Aggressive Behavior, 18, 85-93

Cialdini, R.B., Borden, R.J., Thorne, A., Walker, M.R., Freeman, S., & Sloan, L.R. (1976). Basking in reflected glory: Three (football) field studies. Journal of Personality and Social Psychology, 34, 366-375.

Collegian. (1996). Owners vote to move Oilers to Nashville. Kansas State University: Student Publications Inc.
< > (1996, May 1).

Curtis, J., Loy, J, & Karnilowicz, W. (1986). A comparison of suicide-dip effects of major sport events and civil holidays. Sociology of Sport Journal, 3, 1-14.

Economagic. (2001). Unemployment rate: Houston City, Texas.
< > (2001, January 24).

Durkheim, E. (1897[1951]). Suicide. New York: Free Press.

Fernquist, R.M. (2000). An Aggregate Analysis of Professional Sports, Suicide and Homicide Rates: 30 U.S. Metropolitan Areas, 1971-1990. Aggression and Violent Behavior, 5, 329-341.

Fernquist, R.M. (2001). The 1994-1995 Baseball and Hockey Strikes and Their Impact on Suicide and Homicide Rates in the United States. Archives of Suicide Research, In Press.

Gabennesch, H. (1988). When promises fail: A theory of temporal fluctuations in suicide. Social Forces, 67, 129-145.

Glejser, H. C. (1969). A new test for heteroscedasticity. Journal of the American Statistical Association, 64, 316-323.

Grossberg, L. (1992). “Is there a fan in the house?”: The affective sensibility of fandom. In L. Lewis (Ed.) The adoring audience: Fan culture and popular media. ( pp. 50-65) New York: Routledge.

Johnson, B.D. (1965). Durkheim's one cause of suicide. American Sociological Review, 30, 875-886.

Labor Market Information. (2000). Local area unemployment statistics. 
< > (2000, November 17)

Lester, D. (1994). Patterns of suicide and homicide in America. Commack, NJ: Nova Science Publishers.

Mitrano, J.R. (1999). The “sudden death” of hockey in Hartford: Sports fans and franchise relocation. Sociology of Sport Journal, 16, 134-154. 

New York City Department of Health. Office of Vital Statistics. (Personal correspondence). September 20, 2000.

New York State Department of Health. Bureau of Biometrics. (Personal correspondence). April 22, 2000. 

New York State Department of Labor. Division of Research and Statistics. (Personal correspondence). April 4, 2001.

Ohio Department of Development. (2000). Ohio county profiles: Cuyahoga county.
< > (2000, September 12).

Ohio Department of Health, Center for Public Health Data and Statistics. (Personal correspondence). November 14, 2000

Pooley, J.C. (1980). The sport fan: A social-psychology of misbehaviour. Calgary: University of Calgary: Smith, G.J. (1976). An examination of the phenomenon of sports hero worship. Canadian Journal of Applied Sports Sciences, 4, 259-270.

Stack, S. (1995). Temporal disappointment, homicide and suicide: An analysis of nonwhites and whites. Sociological Focus, 28, 313-328.

Stack, S. (1992). The effect of divorce on suicide in Japan: A time series analysis, 1950-1980. Journal of Marriage and the Family, 54, 327-334.

Sullivan, N.J. (1987). The Dodgers move west. New York: Oxford University Press.

Texas Department of Health. Statistical Services Division, Bureau of Vital Statistics. (Personal correspondence). October 23, 2000

Texas Department of Health. (2000). Texas health data: Population.
< > (2000, October 25)

Texas State Data Center. (2000). Texas population estimates program.
< > (2000, October 25)

Trovato, F. (1998). The Stanley Cup of Hockey and Suicide in Quebec, 1951-1992. Social Forces, 77, 105-127.

Tutko, T.A. (1989). Personality change in the American sport scene. In J.H. Goldstein (Ed.), Sports, games, and play (pp. 111-127). New York: Wiley and Sons.

USA Today. (1995). Modell turns Browns fans blue over move.
< > (1995, November 5).

Wann, D.L. (1996). Seasonal changes in spectators' identification and involvement with and evaluations of college basketball and football teams. The Psychological Record, 46, 201-216.

Wann, D.L., Melnick, M.J., Russell, G.W., & Pease, D.G. (2001). Sport fans: The psychology and social impact of spectators. New York: Routledge.

| current contents | submitting?|

Hosted by the School of Physical Education, University of Otago.

Contact Robyn Jones  for comment.