Aggression and Violent Behavior, Vol. 5, No. 4, pp. 329–341, 2000 Copyright 2000 Elsevier Science Ltd Printed in the USA. All rights reserved 1359-1789/00/$–see front matter
PII S1359-1789(98)00012-3
AN AGGREGATE ANALYSIS OF PROFESSIONAL SPORTS, SUICIDE, AND HOMICIDE RATES: 30 U.S. METROPOLITAN AREAS, 1971–1990 Robert M. Fernquist Central Missouri State University
ABSTRACT. Research on fan behavior in the sociology of sport has discussed the issue of fan behavior at or shortly after the sporting event in detail. Less prevalent in the sociology of sport literature, however, are how successes and failures of sports teams are related to the general public. In this article, suicide and homicide rates are compared to the successes and failures of professional sports teams in 30 metropolitan areas from 1971–1990. Using Gabennesch’s (1988) theory of broken promises as a theoretical framework, statistical analyses reveal that making the playoffs is significantly related to a decline in suicide and homicide rates, while winning championships is significantly related to a decline only for suicide rates. Explanations for these associations are offered. 2000 Elsevier Science Ltd. All rights reserved. KEY WORDS. Suicide, homicide, performance of professional sports teams RECENT SURVEYS REPORT THAT many Americans enjoy professional sports (Gallup, 1992; Iso-Ahola & Hatfield, 1986). Indeed, Gallup (1992, p. 37) finds that 33% of Americans surveyed are either “very interested” or “extremely interested” in following the major professional sports (e.g., baseball, football, basketball, and hockey). Only 25% of respondents were not even “somewhat interested” in professional baseball, football, basketball, or hockey teams. From 1971-1980 A. C. Nielson Co. (1981) reports that onefourth to one-third of households watched at least one game of the World Series, while between 39% and 47% watched the Super Bowl. Furthermore, almost one-third of “the 25 most watched television programs of all time” are Super Bowls (Greendorfer, 1983, p. 1). Sport heroes (and teams) come to personify American values (such as success, freedom, and patriotism; Williams, 1970), and hence become “objects of admiration and emulation” (Leonard, 1980, p. 77). Clearly, however, there are differences in who follows sports more closely; males, for example, are much more likely than females, and younger
Correspondence should be addressed to Robert M. Fernquist, Department of Sociology and Social Work, Central Missouri State University, Wood Hall 203, Warrensburg, MO 64093.
329
330
R. M. Fernquist
people are more likely than older people, to be interested in spectator sports (Gallup, 1992; Pooley, 1980). Although much has been written about fan behavior during or shortly after attending sporting events (Iso-Ahola & Hatfield, 1986; Pooley, 1980; Sloan, 1989; Tutko, 1989), the relationship between the performance of professional sporting teams and annual suicide and homicide rates is the focus herein. The influence of success or failure of professional teams on local suicide and homicide rates in the 30 metropolitan areas is the subject of this research. More specifically, suicide and homicide rates are studied in relation to the successes and failures of professional baseball, basketball (NBA and ABA), football, and hockey teams in these metropolitan areas from 1971–1990. Because sports fans have reacted very strongly to “their” teams’ wins and losses (Sloan, 1989; Tutko, 1989), it is plausible that suicide and homicide rates are related to the performance of professional sport teams.
SPORTS AND SOCIALIZATION Although sports’ fans may hear the phrase “It’s not whether you win or lose, it’s how you play the game,” winning often becomes the most important aspect of their sports interests. Why do fans give such energy to something over which they ultimately have no control? Schafer (1969) argues that sports offer fans a sense of identification, and that the fans feel as if they belong to a social group. It is no surprise, then, that such a strong interest in sports may come to represent an extension of the self and that fans’ attitudes and feelings about themselves may be heavily influenced by how “their team” is doing (Schafer, 1969, p. 33). “Sport involves . . . the celebration of human achievement. It provides fans with a set of organizing principles which give meaning to their” experiences (Edwards, 1973, p. 26). When watching professional sports, fans are often reminded about their own trials and pressures because “athletics . . . excites our dominant personality features” (Tutko, 1989, p. 115). Hence, the team and the fan become one. 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. (Tutko, 1989, p. 115)
Coakley (1990, pp. 146–147), however, argues that “if . . . spectators are concerned with outcomes (of professional sporting events), and if they see sport as a means of proving something to themselves or others, then frustration” leading to destructive behavior is likely to occur if the outcomes are unfavorable. In recent decades, there has been a shift away from, “it’s not whether you win or lose but how you play the game” to, “winning is the only thing” (Tutko, 1989, p. 116). Because the professional sporting arena “crystallizes all of the many values, as well as fantasies, we have, when our team wins we win and our identity is assured . . . ; it is more than a game being played out there—our very personalities are on the line” (Tutko, 1989, pp. 122–123). Whether or not such attachments to professional sports are related to suicide and homicide rates in various metropolitan areas remains to be seen. Research on homicide rates in the general population commonly shows that homicide rates do not significantly vary during, as opposed to before or after, sporting ceremonies such as the last game of the World Series and the Super Bowl (Curtis, Loy, & Karnilowicz, 1986; Lester, 1988; Phillips, 1983). White (1989), however, finds that homicide rates signifi-
Professional Sports, Suicide, and Homicide Rates
331
cantly increased 6 days following a loss in a professional football game for 19 metropolitan areas from 1973 through 1979. White (1989) suggested that, due to the violent nature of professional football, “violent behavior is more likely to be imitated when accompanied by losses and disappointments” (p. 432). Regarding aggressive acts in general, Drake and Pandey (1996) report that fluctuation in violent acts against children the day of, as well as the day following, professional hockey games in St. Louis was unrelated to the performance of the hockey team. On the other hand, White, Katz, and Scarborough (1992) found that, in northern Virginia, women were more likely to be admitted to emergency rooms as trauma victims1 the day of a victory by the Washington Redskins, and were less likely to be admitted 2 days following a Redskins’ loss. White et al. argued that winning increased the number of females admitted to hospitals because “Winning reminds violenceprone viewers that power and aggression work and make them feel good” (p. 167). For suicide rates, Curtis et al. (1986, p. 8) report that, regarding the last game of the World Series and Super Bowl Sunday from 1972–1978, there were significant declines in suicides “on the days just before,” as well as the day of, the sporting event. For the days immediately following the sporting events, there were significant increases in suicide. Lester (1988), however, finds no significant change in suicide rates on Super Bowl Sunday or on the last game of the World Series from 1972 to 1984 compared to suicides and homicides 1 week before and 1 week after these events. Using mortality data specific to the city in which the sporting event occurred, Phillips and Feldman (1973) find that total mortality rates decline “before the Olympic Games . . . in the cities that” host the Olympics. If professional sporting events have taken a prominent place in the lives of many Americans, as Loy, McPherson, and Kenyon (1978) argue, then more research needs to be done on the association between professional sporting events and suicide and homicide rates in metropolitan areas. Pope (1998), while discussing Durkheim’s (1961) discourse on religious symbols and ceremonies, argues that sporting events promote integration among those who attend the event. Using collegiate football as an example of how sports promotes integration among the fans, he writes that During the game fans energetically take advantage of numerous opportunities to collectively participate in common activities demonstrating their shared beliefs, emotions, and togetherness . . . . In short, sporting events exemplify the conditions of religious ritual: high rates of group interaction, focus on sacred symbols, and collective ritual behavior symbolizing group membership and strengthening shared beliefs, values, aspirations, and emotions. (pp. 53–54)
For this analysis, however, it is hypothesized that such integration, interconnectedness, and high emotional intensity are shared by the majority of people who watch or listen to sporting events. Although data used herein will not permit individual-level references, if suicide and homicide rates are related to the performance of sports teams, this would suggest the above hypothesis is plausible.
SPORTS, TIME-RELATED PHENOMENA, AND BROKEN PROMISES Stack (1995, p. 313) writes that “the impact of time-related phenomena on (suicide and homicide) has been a relatively neglected area of research.” He further writes that temporal
1 White, Katz, and Scarborough (1992) examined admittance for gun shot wounds, stabbings, assaults, falls, being struck by objects, and lacerations.
332
R. M. Fernquist
effects need to be studied on both suicide and homicide rates because both rates are impacted by temporal phenomena. To this end, Gabennesch’s (1988) theory of broken promises is discussed below. Focusing his remarks on suicide, Gabennesch (1988) writes that The suicidal persons affective state can be adversely influenced by circumstances which tend to promote aspiration or expectation for feeling better . . . . The critical point is that a dysphoric mood is the result of the negative discrepancy between how one feels and how one expects (hopes) to feel . . . . In sum, a temporal broken-promise effect can develop from the elevated sense of expectancy implicitly occasioned by . . . a positively valued event. (pp. 138–139)
Therefore, when expectations are not met, serious disappointment may result, and such disappointment could2 lead to suicide (Gabennesch, 1988) with the passage of time. Gabennesch further maintains that suicide is likely to occur due to “conditions or events which induce psychological misery” (p. 142); this certainly is relevant to the arena of professional sports since persons who expect “their” team to do well may become extremely disappointed when the “promise” of winning is broken. Stack (1995) argues that Gabennesch’s theory of broken promises can also be applied to homicide. Stack (1995) writes that because “a central category in Durkheim’s . . . short discussion of anomic homicide was disappointment . . . , homicide (could also) emanate from disappointments or broken promises” (pp. 314–315). In Durkheim’s (1897/1951, p. 357) own words, he argues 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). To broaden the understanding of how temporal variables are related to both suicide and homicide rates, Gabennesch’s (1988) theory of broken promises is used as the theoretical framework for this analysis. Since failures of sports teams (especially repeated failures) are indicative of broken promises, it is hypothesized that failures of professional sports teams are positively related to both suicide and homicide rates.3 Furthermore, because not winning a championship should be a more serious “broken promise” than not making the playoffs, it is hypothesized that championship status will have a stronger impact on suicide and homicide rates than simply making the playoffs. Because success in professional sports is often measured by making the playoffs or winning a championship (Tutko, 1989, p. 116), the following comparisons are made: (1) metropolitan areas that make the playoffs frequently versus metropolitan areas that do not; (2) metropolitan areas with at least one championship versus metropolitan areas without any championships. This latter comparison will be elaborated upon to compare “Big” championship areas to all other areas. Relevant control variables included in these analyses are divorce rates, employment rates, and region of the country. Sociological research indicates that divorce rates (Fernquist, 1997; Lester, 1993; Stack & Wasserman, 1993) and unemployment rates (Lester, 1994b) are positively associated with rates of suicide and homicide in America because
2
Gabennesch (1988) says that “A broken-promise effect is no more likely to have a socially uniform impact than is any other independent variable” (p. 142). 3 Although external aggression increased during recent championship celebrations (e.g., Chicago Bulls 1991–1993, Dallas Cowboys 1992–1993), we hypothesize that the broken promise effect will have an even greater tendency to increase external aggression, at least in the form of increased homicide rates.
Professional Sports, Suicide, and Homicide Rates
333
of weakened social integration. Messner and Golden (1992) find no significant difference in homicide rates according to the region of the country, while Unnithan, Huff-Corzine, Corzine, and Whitt (1994) find no relation between region and suicide and homicide rates. Other research, however, shows that the South tends to have higher homicide rates than other regions of the country (e.g., see Land, McCall, & Cohen, 1990; Lester, 1986). Lester (1986) reports no significant variation in total suicide rates according to the region of the country. Because “the strength of the temporal broken-promise effect varies across demographic categories” (Gabennesch, 1988, p. 142), race, gender, and age structure are also controlled. In America, Whites, males, and older persons have higher suicide rates than non-Whites, females and younger persons; homicide rates are higher among non-Whites, males, and younger persons compared to Whites, females, and older persons (Fernquist, 1996; Lester, 1994a; Stack, 1995; U.S. Department of Health and Human Services, 1995).
METHODS AND MEASUREMENT All variables are taken from component counties of 30 Metropolitan Statistical Areas (MSAs) from 1971–1990, using the U.S. Bureau of the Census’s (1974) definition of MSA component counties.4 Only those metropolitan areas which had at least one professional sports team for the entire 20-year period are examined. Figures on the number of suicides and homicides (U.S. Department of Health, Education, and Welfare, 1974–1979, Vol. II, Part A; U.S. Department of Health and Human Services, 1980–1993, Vol. II, Part A) are divided by population estimates in these same areas (U.S. Bureau of the Census, 1972a, 1984, 1994) to obtain rates per 100,000 population. The average number of times a metropolitan area had at least one team make the playoffs from 1971–1990 is 13.4. To measure playoff status, the metropolitan areas are dichotomized into two groups—those areas which had at least one team make the playoffs in 14 or more of the 20 years are coded as 1, and those with 13 or fewer playoff appearances are coded as 0. To measure championship status, the metropolitan areas are dichotomized into two groups—those with at least one championship are coded as 1 and those with no championships are coded as 0. Because five metropolitan areas (e.g., New York, Boston, Pittsburgh, Los Angeles, and San Francisco/Oakland) account for 59% of all championships won from 1971–1990, these five areas as are coded as 1 and all other areas as 0 to see if differences in suicide and homicide rates exist between big championship areas and all other areas (this variable is called Big Championship Status). Data for both making the playoffs and the number of championships won are obtained from Johnson (1972–1992). Data on the number of divorces are from the U.S. Department of Health, Education, and Welfare (1974–1979, Vol. III) and the U.S. Department of Health and Human Services (1980–1996, Vol. III). Because the most recent published divorce rates for metropolitan areas cover only 1988, 1988 rates are used for 1989–1990. Data on the total number of
4 Divorce and unemployment statistics were not available for each MSA, but were present at the county level. Therefore, the individual counties in each MSA were summed to yield the MSA divorce and employment rates. The exceptions to this were the demographic variables and population estimates, where data were taken directly from the MSAs. However, very few changes in component counties for the metropolitan areas in this study from 1970–1990 occurred (U.S. Bureau of the Census, 1972b, Appendix D; Slater & Hall, 1993; Appendix B).
334
R. M. Fernquist
nonagricultural employees are from the U.S. Department of Labor (1989, 1994). Because these data are available only back to 1972, the 1972 rates are also used for 1971. Data for the following two areas were also not available: Seattle from 1972–1981 and Buffalo 1972–1974. To include these two areas in the aggregate analyses, the latest data available are used for all preceding years (i.e., 1975 data for Buffalo are used for 1971–1974). Both divorce and employment data are divided by population estimates for each area (U.S. Bureau of the Census, 1972a, 1984, 1994) to obtain rates per 1,000 population. The measure of region herein is a dummy variable with those areas located in the South, as defined by the U.S. Bureau of the Census (1996; e.g., Atlanta, Dallas, Houston, Miami, New Orleans, and Washington, DC) coded as 1 and all other areas coded as 0. Huff-Corzine, Corzine, and Moore (1986, pp. 909–910) argue that measuring Southern culture as a dummy variable is flawed because regional factors do not end at state boundaries. However, Land et al. (1990) find that the dummy variable measure of the South is very highly correlated (.88 or higher) with other measures of Southern culture such as Gastil’s (1971) index of southernness. For the demographic variables, race is measured as the percent of the population that is non-White; sex is measured as the percent of the population that is female; age is measured as the percent of the population that is 65 years of age and older (U.S. Bureau of the Census, 1972b, Table 3; U.S. Bureau of the Census, 1986, Table A; Slater & Hall, 1993, Table B). Only census data are used for the demographic variables and population figures because noncensus data yield very different percentages. To obtain data on these variables for each year, estimates were extrapolated from the existing data using an assumption of linear change over time. For example, the percentage of the population aged 65 and over in New York in 1970 was 10.9%, and it was 12.2% in 1980. Therefore, the 1971 percentage was estimated as 11.0%, the 1972 percentage was estimated as 11.1%, and so on through 1979. A modified generalized least squares approach (MGLS) is used for the analyses of the 30 metropolitan areas because MGLS is appropriate for a pooled cross-sectional time series research design (Cutright & Briggs, 1995). Because there are 30 different metropolitan areas in this sample, each with annual observations from 1971–1990, the sample size is 600 (e.g., 30 ⫻ 20 ⫽ 600). MGLS estimation involves a weighted transformation of the dependent variable which corrects for both cross-sectional heteroskedasticity and firstorder autocorrelation. See Kmenta (1986) for more detail on MGLS estimation. The traditional R 2 with OLS techniques is not a valid indicator of explained variance when MGLS techniques are employed (Greene, 1990), and so an alternative measure of explained variance (Buse R 2) is used as described by Judge, Griffiths, Hill, Lutkepohl, and Lee (1985). Estimated rho values in MGLS estimation show how much statistical efficiency is gained by using MGLS instead of OLS estimation. The larger the rho, the more efficient the MGLS estimates. For all models in Table 2, the estimated rho values show that statistical efficiency was gained by using MGLS estimation. To check for problematic collinearity, each independent variable was regressed on all others simultaneously, and no R2 was above .49 for all models in Table 2. This suggests that problematic collinearity is not present in these data.
RESULTS Bivariate relationships among all variables, as well as means and standard deviations, are shown in the Appendix. Table 1 shows homicide rates, suicide rates, the number of
Professional Sports, Suicide, and Homicide Rates
335
TABLE 1. Homicide Rates, Suicide Rates, and Number of Championships Won for 30 Metropolitan Areas (1971–1990) No. Times Making Playoffs
Homicide Rate
Suicide Rate
Championships
New Orleans Miami Houston New York Detroit Los Angeles Dallas Atlanta Chicago St. Louis Baltimore Cleveland Kansas City Washington, DC Philadelphia Oakland/San Francisco Phoenix Indianapolis Denver San Diego Cincinnati Milwaukee Boston Buffalo Seattle Portland Anaheim Pittsburgh Minneapolis/St. Paul Green Bay
23.42 22.68 21.69 19.34 18.23 17.18 16.39 15.26 14.40 14.35 13.98 13.81 11.71 11.54 11.02 9.89 9.16 8.48 7.65 7.31 6.88 6.23 6.01 5.73 5.65 5.35 5.24 4.60 3.40 1.77
13.39 18.23 14.31 7.54 12.38 15.19 13.20 12.78 9.81 11.15 10.90 13.16 13.42 10.80 11.27 18.00 17.28 12.54 18.89 15.48 12.59 13.56 10.76 9.29 14.61 14.30 13.14 11.65 11.14 12.28
0 2 0 10 3 8 2 0 1 1 2 0 1 3 4 11 0 2 0 0 3 1 6 0 1 1 0 6 1 0
1 12 10 20 15 20 18 14 20 17 9 12 9 15 19 17 11 9 15 5 11 17 20 15 12 13 8 16 20 3
All areas
11.28
13.12
69
13.4
Metropolitan Area
Note. Cell values are based on 20 cases; column totals are based on 600 cases.
championships won, and the number of times at least one team made the playoffs for each metropolitan area during the 1971–1990 period. The highest homicides rates are in New Orleans, Miami, and Houston, while the highest suicide rates are in Denver, Miami, and Oakland/San Francisco. With the exception Oakland/San Francisco, then, the areas with the highest homicide and suicide rates have either few or no championships during the 20-year period. Bivariate correlations (see the Appendix) show that there is no pattern of association between homicide rates and team performance. Suicide rates, however, are inversely related to playoff status, championship status, and big championship status. Unstandardized MGLS estimates (top panel of Table 2) reveal that playoff status and championship status are negatively related to suicide rates (b ⫽ ⫺.752, p ⫽ .006 and ⫺.792, p ⫽ .013, respectively; standardized coefficients are in parentheses). Playoff status
336
R. M. Fernquist
TABLE 2. MGLS Estimates of Suicide and Homicide Rates in 30 Metropolitan Areas (1971–1990) Dependent Variable Suicide Rate Independent Variable
Metric
Standardized
Employment rate Divorce rate Playoff status Championship status South Percent non-White Percent female Percent 65⫹
⫺.038* .702*** ⫺.752** ⫺.792* .689 ⫺.029 ⫺.786*** .360***
(⫺.076) ( .354) (⫺.123) (⫺.122) ( .090) (⫺.084) (⫺.210) ( .238)
Constant Buse R2 Estimated rho
49.189 .354 .778
Employment rate Divorce rate Playoff status Big championship status South Percent non-White Percent female Percent 65⫹
⫺.051** .616*** ⫺.689* ⫺.910 1.810*** ⫺.063** ⫺.811*** .425***
Constant Buse R2 Estimated rho
50.679 .257 .780
Homicide Rate Metric ⫺.039 .535*** ⫺.706 ⫺.132 3.121*** .492*** 2.018*** ⫺.366**
Standardized (⫺.037) ( .130) (⫺.056) (⫺.010) ( .197) ( .688) ( .260) (⫺.117)
⫺98.733 .670 .865 (⫺.101) ( .311) (⫺.113) (⫺.111) ( .237) (⫺.182) (⫺.217) ( .281)
⫺.047* .567*** ⫺1.046* ⫺1.308 3.131*** .493*** 2.024*** ⫺.408**
(⫺.045) ( .138) (⫺.082) ( .077) ( .198) ( .689) ( .260) (⫺.130)
⫺98.446 .677 .864
Note. Standardized coefficients are in parentheses. *p ⬍ .03. **p ⬍ .01. ***p ⬍ .001.
and championship status, however, are both unrelated to homicide rates in the top panel. For the control variables in the MGLS suicide model, divorce rates and percent 65 and over are related to increased suicide rates, while employment rates and percent female are related to a decline in suicide rates. For the homicide model, divorce rates, Southern areas, percent non-White, and percent female are related to increased homicide rates. Percent 65 and older is inversely related to homicide rates. Taken as a whole, these variables explain variation in homicide rates better than in suicide rates. When metropolitan areas are dichotomized into areas with six or more championships versus those with no championships (bottom panel, Table 2), big championship status is not significantly related to either suicide rates (b ⫽ ⫺.910, p ⫽ .065) or homicide rates (b ⫽ ⫺1.308, p ⫽ .070). Playoff status remains inversely related to suicide rates, but it is also inversely related to homicide rates in the bottom panel (b ⫽ ⫺1.046, p ⫽ .020). All other associations remain as they were in the top panel of Table 2, except that region and percent non-White become significantly related to suicide rates (b ⫽ 1.810, p ⫽ .000
Professional Sports, Suicide, and Homicide Rates
337
and b ⫽ ⫺.063, p ⫽ .001, respectively), whereas employment status becomes significantly related to homicide rates (b ⫽ ⫺.047, p ⫽ .021). If the standardized coefficients are used to compare the relative strength one independent variable has on suicide and homicide rates versus all other independent variables, then divorce rates are most strongly related to suicide rates and percent non-White is most strongly related to homicide rates (this is true for both panels in Table 2). For championship areas only, we intended to compare suicide and homicide rates for the years a championship was won versus 1 year before, as well as 1 year after, the championship year. However, a regression that was free from both heteroskedasticity and autocorrelation simultaneously could not be produced.5 In lieu of regression analyses, a difference of means test reveals that there are no significant differences in suicide and homicide rates during championship years compared to suicide and homicide rates the year before (t ⫽ 0.12, t ⫽ 1.05, respectively) or the year after (t ⫽ 0.76, t ⫽ 0.94, respectively) the championship year.
CONCLUSION Data from 30 metropolitan areas from 1971–1990 show some support Gabennesch’s (1988) theory of broken promises when his theory is applied to the performance of professional sports teams. Compared to metropolitan areas with no championships between 1971 and 1990, metropolitan areas with at least one championship had lower suicide rates. Winning a championship, therefore, is indicative of keeping promises, at least as far as suicide rates are concerned. Playoff status is also indicative of keeping promises since areas with teams who had an above average number of playoff appearances had lower suicide and homicide rates compared to below average areas. From Table 2, however, we see that playoff status is significantly related to the dependent variables in three of the four models, whereas championship status is significantly related to a dependent variable in only one of the four models. Therefore, even though both playoff status and championship status have significant, inverse associations with suicide and homicide rates, the hypothesis that championship status, as opposed to playoff status, would have a stronger impact on suicide and homicide rates is not supported. Further tests of this hypothesis are needed, however, before offering reasons why this hypothesis was not supported. Below, comments are directed toward playoff status since it had more significant associations with the dependent variables. Because playoff status is inversely related to both suicide and homicide rates, this implies that when teams reach the playoffs on a regular basis, expectations that teams do well are met (Gabennesch, 1988; Stack, 1995).6 The mechanism by which Gabennesch’s theory of broken promises is tied to suicide rates, homicide rates, and the performance of professional sports teams is the strong identification individuals have with professional sports teams. Because affiliation with sports teams can come to represent an extension of the self, and seeing that fans’ attitudes and feelings about themselves may be heavily
5 To use MGLS estimation, the statistical package used herein (Shazam) requires that there be an equal number of observations for each cross section—this was not the case when comparing championship years to the years before and after the championship. As a result, OLS estimation was used, but OLS could not produce an acceptable model. 6 Although Robinson (1950) has shown that individual-level references cannot be made from aggregatelevel data, Gabennesch’s theory is still applicable to the above statistical associations.
338
R. M. Fernquist
influenced by how “their team” is doing (Schafer, 1969), a strong identification with professional sports teams can occur. What drives such a strong identification with professional sports teams? Professional athletes possess two qualities that are cherished by many in society—cherished enough to promote a strong sense of identification with professional sports teams: Extraordinary athletic talent and exorbitant incomes. Admiring these two qualities as many individuals do, broken promises in the professional sports world have a profound impact on suicide and homicide rates because individuals’ identification with the team (or player) is threatened. Contrary to Gabennesch’s (1988) arguments, however, it is not necessary for people to first be in a suicidal (or homicidal) state for the broken promise effect to have an adverse impact on them (at least as the broken promise effect has been applied herein). Since identification with professional sports teams can be a very powerful influence in individuals’ lives (Cialdini et al., 1976; Schafer, 1969; Sloan, 1989; Tutko, 1989), a broken promise in professional sports in and of itself is powerful enough to promote an adverse effect on those individuals who strongly identify with professional sports teams. The intensity of fan identification with sports teams, of course, does vary (Wann, 1996). Some fans strongly identify with teams on a year-round basis, others are fair-weather fans, and still others have minimal identification with teams. Therefore, it is plausible that those persons who would take their own lives, or someone else’s life, due to (in some part) the poor performance of professional sports teams are very strongly identified with the teams, especially since fans who are strongly identified with teams have difficulty distancing themselves from “their” teams’ failures (Branscombe & Wann, 1992; Wann & Branscombe, 1990). Just why aggression is sometimes directed inward and sometimes directed against others, however, is as yet unclear. A more refined test of broken promises would be to compare suicide and homicide rates of the championship team with the team that loses in the finals; the data used herein, however, do not permit such a test. Monthly data on suicide and homicide rates would be more appropriate to ascertain the impact of winning a championship versus losing in the finals in the same year. Such research could also further determine if playoff status has a stronger impact on suicide and homicide rates than championship status. As Eitzen and Sage (1993) point out, a deeply rooted “heritage and culture of violence . . . underlie much of the violence in contemporary North American society” (p. 167). Apparently, this heritage extends to the relationship between suicide rates, homicide rates, and professional sports. Further studies, both at the individual and aggregate-level, could yield added knowledge to how the performance of professional sports teams is related to suicide and homicide rates. Further research would do well to examine other aspects of the relationship between suicide and homicide rates and the performance of professional sports teams, such as the impact of separate sports (e.g., baseball, football, etc.) on suicide and homicide rates; the effect that teams moving from one area to another has on suicide and homicide rates; and the effect of strife in professional sports on suicide and homicide rates. Acknowledgments—Thanks are expressed to Whitney Pope, Phillips Cutright, David Lester, Bijou Lester, and Steven Stack for comments on earlier versions of this article. I also thank Whitney Pope for the initial idea of this research.
REFERENCES A. C. Nielson Co. (1981). Let’s look at sports, 1980. Northbrook, IL: Author. Branscombe, N. R., & Wann, D. L. (1992). Role of identification with a group, arousal, categorization processes, and self-esteem in sports spectator aggression. Human Relations, 45, 1013–1033.
Professional Sports, Suicide, and Homicide Rates
339
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. Coakley, J. J. (1990). Sport in society. St. Louis: Times Mirror/Mosby. Curtis, J., Loy, J., & Karnilowicz, W. (1986). A comparison of suicide-dip effects of major sports events and civil holidays. Sociology of Sport Journal, 3, 1–14. Cutright, P., & Briggs, C. M. (1995). Structural and cultural determinants of adult homicide in developed countries: Age and gender specific rates, 1955-1989. Sociological Focus, 28, 221–243. Drake, B., & Pandey, S. (1996). Do child abuse rates increase on those days on which professional sporting events are held? Journal of Family Violence, 11, 205–218. Durkheim, E. (1951). Suicide: A study in sociology (J. A. Spaulding & G. Simpson, Trans.). New York: Free Press. (Original work published 1897) Durkheim, E. (1961). The elementary forms of the religious life (J. W. Swain, Trans.). New York: Collier. (Original work published 1915). Edwards, H. (1973). Sociology of sport. Homewood, IL: Dorsey. Eitzen, D. S., & Sage, G. H. (1993). Sociology of North American sport. Madison, WI: Brown and Benchmark. Fernquist, R. M. (1996). Social and psychological determinants of gender and age-specific suicide rates in developed countries. Ann Arbor, MI: UMI Dissertation Services. Fernquist, R. M. (1997). The 1994–1995 baseball and hockey strikes and their impact on suicide and homicide rates in the United States. Unpublished manuscript. Gabennesch, H. (1988). When promises fail: A theory of temporal fluctuations in suicide. Social Forces, 67, 129–145. Gallup, G. (1992, October). The Gallup poll monthly. Princeton, NJ: The Gallup Organization. Gastil, R. D. (1971). Homicide and regional culture of violence. American Sociological Review, 36, 412–427. Greendorfer, S. L. (1983). Sport and the mass media: General overview. Arena Overview, 7, 1–6. Greene, W. (1990). Economic analysis. New York: Macmillan. Huff-Corzine, L., Corzine, J., & Moore, D. C. (1986). Southern exposure: Deciphering the South’s influence on homicide rates. Social Forces, 64, 906–924. Iso-Ahola, S., & Hatfield, B. (1986). Psychology of sports: A social psychological approach. Dubuque, IA: Brown. Johnson, O. (Ed.). (1972–1992). Information please almanac. Boston: Houghton-Mifflin. Judge, G. G., Griffiths, W. E., Hill, R. C., Lutkepohl, H., & Lee, T. C. (1985). The theory and practice of econometrics (2nd ed., rev.). New York: Wiley. Kmenta, J. (1986). Elements of econometrics. New York: Macmillan. Land, K. C., McCall, P. L., & Cohen, L. E. (1990). Structural covariates of homicide rates: Are there any invariances across time and space? American Journal of Sociology, 95, 922–963. Leonard, W. M., III. (1980). A sociological perspective of sport. Minneapolis, MN: Burgess. Lester, D. (1986). Southern subculture, personal violence (suicide and homicide), and firearms. Omega, 17, 183–186. Lester, D. (1988). Suicide and homicide during major sports events 1972–1984: Comment on Curtis, Loy, and Karnilowicz. Sociology of Sport Journal, 5, 285. Lester, D. (1993). Marital integration, suicide, and homicide. Psychological Reports, 73, 1354. Lester, D. (1994a). Patterns of suicide and homicide in America. Commack, NJ: Nova Science. Lester, D. (1994b). Suicide and unemployment: A monthly analysis. Psychological Reports, 75, 602. Loy, J. W., McPherson, B. D., & Kenyon, G. (1978). Sport and social systems. Reading, MA: Addison-Wesley. Messner, S. F., & Golden, R. M. (1992). Racial inequality and racially disaggregated homicide rates: An assessment of alternative theoretical explanations. Criminology, 30, 421–445. Phillips, D. P. (1983). The impact of mass media violence on U.S. homicides. American Sociological Review, 48, 560–568. Phillips, D. P., & Feldman, K. A. (1973). A dip in deaths before ceremonial occasions: Some relationships between social integration and mortality. American Sociological Review, 38, 678–696. Pooley, J. C. (1980). The sport fan: A social-psychology of misbehaviour. Ottawa: Cahper. Pope, W. (1998). Emile Durkheim. In R. Stones (Ed.), Key sociological thinkers. New York: Macmillan. Robinson, W. A. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351–357. Schafer, W. E. (1969). Some social sources and consequences of interscholastic athletics. In G. S. Kenyon (Ed.), Aspects of contemporary sociology of sport (pp. 29–44). Chicago: The Athletic Institute. Slater, C. M., & Hall, G. E. (1993). 1993 county and city extra. Lanham, MD: Bernan. Sloan, L. R. (1989). The motives of sports fans. In J. H. Goldstein (Ed.), Sports, games, and play (pp. 175–240). New York: Wiley. Stack, S. (1995). Temporal disappointment, homicide and suicide: An analysis of nonwhites and whites. Sociological Focus, 28, 313–328. Stack, S., & Wasserman, I. (1993). Martial status, alcohol consumption, and suicide: An analysis of national data. Journal of Marriage and the Family, 55, 1018–1024.
340
R. M. Fernquist
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. Unnithan, N. P., Huff-Corzine, L., Corzine, J., & Whitt, H. P. (1994). The currents of lethal violence. New York: State University of New York Press. U.S. Bureau of the Census. (1972a). County and city data book. Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1972b). Statistical abstract of the United States. Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1974). Map GE-50 (No. 55). Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1984). Statistical abstract of the United States. Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1986). County and city data book. Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1994). Statistical abstract of the United States. Washington, DC: U.S. Government Printing Office. U.S. Bureau of the Census. (1996). Statistical abstract of the United States. Washington, DC: U.S. Government Printing Office. U.S. Department of Health and Human Services. (1980–1993). Vital statistics of the United States (Vol. II, Part A). Hyattsville, MD: U.S. Government Printing Office. U.S. Department of Health and Human Services. (1980–1996). Vital statistics of the United States (Vol. III). Hyattsville, MD: U.S. Government Printing Office. U.S. Department of Health and Human Services. (1995). Health in the United States, 1994. Washington, DC: U.S. Government Printing Office. U.S. Department of Health, Education, and Welfare. (1974–1979). Vital statistics of the United States (Vol. II, Part A). Hyattsville, MD: U.S. Government Printing Office. U.S. Department of Health, Education, and Welfare. (1974–1979). Vital statistics of the United States (Vol. III). Hyattsville, MD: U.S. Government Printing Office. U.S. Department of Labor. (1989). Employment hours and earnings, states and areas 1972–1987. Washington, DC: U.S. Government Printing Office. U.S. Department of Labor. (1994). Employment hours and earnings, states and areas 1988–1994. Washington, DC: U.S. Government Printing Office. 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., & Branscombe, N. R. (1990). Die-hard and fair-weather fans: Effects of identification on BIRGing and CORFing tendencies. Journal of Sport and Social Issues, 14, 103–117. White, G. F. (1989). Media and violence: The case of professional football championship games. Aggressive Behavior, 15, 423–433. White, G. F., Katz, J., & Scarborough, K. E. (1992). The impact of professional football games upon violent assaults on women. Victims and Violence, 7, 157–171. Williams, R. M. (1970). American society: A sociological interpretation. New York: Alfred Knopf.
341
1.00 .02 ⫺.04 .13 .01 .02 .01 .57 .75 .17 ⫺.09 11.3 6.3
Homicide rate (1) Suicide rate (2) Employment rate (3) Divorce rate (4) Playoff status (5) Championship status (6) Big championship status (7) South (8) Percent non-White (9) Percent female (10) Percent 65⫹ (11)
M SD
13.1 3.1
1.00 ⫺.05 .56 ⫺.27 ⫺.21 ⫺.07 .11 ⫺.07 ⫺.37 ⫺.06
(2)
Note. If r is between .07 and .09, p ⬍ .05; if r ⬎ .09, p ⬍ .01.
(1)
Variable
43.9 6.0
1.00 ⫺.04 .08 .15 .10 .04 .14 .10 .09
(3)
4.9 1.5
1.00 ⫺.37 ⫺.15 ⫺.24 .24 ⫺.02 ⫺.40 ⫺.23
(4)
0.5 0.5
1.00 .33 .42 ⫺.03 .13 .28 .03
(5)
0.7 0.5
1.00 .32 ⫺.18 .04 .37 .29
(6)
0.2 0.4
1.00 ⫺.22 .09 .18 .32
(7)
APPENDIX Intercorrelations Among Variables Used in Analyses (N ⫽ 600)
0.2 0.4
1.00 .51 .02 ⫺.35
(8)
17.9 8.9
1.00 ⫺.01 ⫺.15
(9)
51.5 0.8
1.00 .50
(10)
10.4 2.0
1.00
(11)