The Long Shadow of Rivalry: Rivalry Motivates Performance Today and Tomorrow

Abstract

Research has established that competing head to head against a rival boosts motivation and performance. The present research investigated whether rivalry can affect performance over time and in contests without rivals. We examined the long-term effects of rivalry through archival analyses of postseason performance in multiple high-stakes sports contexts: National Collegiate Athletic Association (NCAA) Division I Men’s Basketball and the major U.S. professional sports leagues: National Basketball Association (NBA), National Football League (NFL), Major League Baseball (MLB), and National Hockey League (NHL). Econometric analyses revealed that postseason performance of a focal team’s rival in year N predicted that focal team’s postseason performance in year N + 1. Follow-up analyses suggested that the performance boost was especially pronounced when one’s rival won the previous tournament. These results establish that rivalry has a long shadow: A rival team’s success exerts such a powerful motivational force that it drives performance outside of direct competition with one’s rival and even after a significant delay.

Keywords

rivalry competition goal setting motivation interpersonal interaction intergroup interaction open data

Duke versus University of North Carolina (UNC) is widely considered the greatest rivalry in college basketball (Novak, 2013; Paine, 2017). Combined, these teams have won nine national titles over the last 27 seasons. Even more remarkably, titles on “Tobacco Road” tend to appear in back-to-back years (Galinsky & Schweitzer, 2015): Duke won the title in 1992, UNC in 1993; UNC held the trophy in 2009, Duke in 2010; Duke cut down the nets in 2015, and then UNC made the title game in 2016 and won it all in 2017. Many people recognized that UNC’s 2017 win was not only motivated by their last-second loss in 2016 but also by Duke’s 2015 title. Consider these quotes from Sports Illustrated (Rosenberg, 2017): “People who focus on Villanova’s winning the 2016 championship miss the point. Sure, that motivated the Heels. But so did the ’15 championship, which was won by—excuse our language—Duke” (para. 13). UNC head coach Roy Williams echoed this sentiment: “They [UNC Juniors] were jealous after what Duke did their freshman year . . . And I think it motivated . . . them to work more” (para. 15).

This example suggests two powerful insights regarding rivalry. First, it can foster increased motivation and performance not just today but over the long term. Second, since National Collegiate Athletic Association (NCAA) rivals rarely meet in the tournament, a rival’s performance can motivate a focal team outside of head-to-head competition with the rival. In fact, UNC and Duke have never played each other in the NCAA tournament, yet each team’s tournament success seems to drive the other’s future performance.

The current research explored whether rivalry has a long motivational shadow. Although past research has found that rivalry increases motivation in head-to-head matches with one’s rival (Kilduff, 2014; Kilduff, Elfenbein, & Staw, 2010), we propose that rivalry is so powerful that its effects can leap beyond the confines of head-to-head matches, motivating performance even when rivals are not directly competing. For example, even when Larry Bird was not competing head to head with his rival, Magic Johnson, Magic’s performance motivated Larry: “The first thing I would do every morning was look at the box scores to see what Magic did. I didn’t care about anything else” (Neely, 2012, para. 17).

Additionally, we propose that rivalry can motivate performance over long time periods. Past experimental manipulations of rivalry have typically assessed consequences within a short time horizon (Converse & Reinhard, 2016; Kilduff & Galinsky, 2017), and archival analyses have looked at performance only in one-off contests between rivals (Kilduff, 2014; Kilduff et al., 2010). However, we believe that a rival’s performance, especially in highly salient contexts, will so deeply pervade an actor’s psyche that it will influence his or her motivation and performance not only today but also tomorrow.

The Long Shadow of Rivalry

The existing research on rivalry shows that for both individuals and teams, effort is greater in competitions against rivals as opposed to nonrivals (Kilduff, 2014; Kilduff et al., 2010). Contests against rivals boost motivation by increasing the psychological importance of that competition (Kilduff, Galinsky, Gallo, & Reade, 2016; To, Kilduff, Ordonez, & Schweitzer, 2017). However, this work has explored the effects of rivalry only in head-to-head competitions among rivals. Here, we tested the idea that the strong performance of one’s rival provides more long-term and generalized motivation, boosting performance months later and outside of direct competition with one’s rival.

This prediction is based in the goal-setting literature (Heath, Larrick, & Wu, 1999; Locke & Latham, 1990). We suggest that rivals’ performance will serve as an input into focal actors’ goals and aspirations, thus affecting future performance. Because rivals care strongly about their status vis-à-vis each other (Kilduff et al., 2016), a rival’s performance may serve as a particularly salient and specific goal. Indeed, specific stretch goals are more motivating and beneficial to performance than vague or less ambitious goals, and they can even motivate behavior over the long term (Harackiewicz, Barron, Tauer, & Elliot, 2002; Locke & Latham, 1990). Thus, we predicted that a rival’s strong performance at time N would motivate a focal actor during time N + 1, resulting in increased future performance.

Overview of the Present Research

To test whether rivalry has a long shadow, we examined postseason performance across multiple sports contexts: the NCAA Division I Men’s Basketball Tournament (“the tournament”) and the Big Four U.S. professional sports leagues’ playoffs (“the playoffs”): National Basketball Association (NBA), National Football League (NFL), Major League Baseball (MLB), and National Hockey League (NHL). We hypothesized that a rival’s performance in a given postseason (year N) would predict the focal team’s performance in the subsequent postseason (year N + 1). Further, we explored whether exceptional rival achievements in year N—that is, winning the championship—have particularly strong effects on the focal team’s performance in year N + 1.

By examining a range of sports at both the professional and amateur levels, we sought to show that the long shadow of rivalry is generalizable across competitive contexts. Importantly, postseason performance may offer a conservative test of our hypothesis. Because the objective stakes are so high, motivation should be high for all teams, regardless of how their rival performed in the prior year. This uniformly high level of motivation might crowd out the psychological force of rivalry. If rivalry adds an additional motivational boost on top of the already intense postseason stakes, it would demonstrate how powerful its shadow is.

Study 1: NCAA Division I Men’s Basketball Tournament

Study 1 explored the long shadow of rivalry in the NCAA Division I Men’s Basketball Tournament. We tested whether the tournament performance of a team’s rival in year N predicted that team’s tournament performance in year N + 1.

Method

Study context

The NCAA Division I Men’s Basketball Tournament is a single-elimination tournament involving 64¹ teams. This tournament is referred to as “March Madness” because of the excitement it generates. Indeed, it is one of the most watched events in America; in 2017, 10 million viewers tuned into each game, on average, and more than 20 million viewers watched the championship game (NCAA.com, 2017a, 2017b). Industry experts estimated that $10.4 billion would be wagered throughout the 2017 tournament, roughly double that placed on the Super Bowl (Heitner, 2017; Purdum, 2017). Further, given that the NBA welcomes only 60 new players through the draft every year, March Madness offers the biggest, and often final, stage of these competitors’ athletic careers.

We proposed that tournament performance is a particularly salient outcome on which rivals are likely to compare themselves for several reasons. First, among universities with significant basketball programs, tournament performance represents the primary measure of success, so it should naturally be a basis of comparison between rivals. Second, tournament performance is highly visible; the tournament is watched by millions of people and covered nonstop by the media. Rivals cannot avoid becoming acutely aware of one another’s performance. Third, tournament performance comes at the end of the season, and from a psychological standpoint, endings have a greater impact than events earlier in a sequence (Kahneman, Fredrickson, Schreiber, & Redelmeier, 1993; Redelmeier & Kahneman, 1996). Indeed, the extended time gap between the end of the tournament and the start of the next season (more than 6 months) gives teams plenty of time to ruminate on their rival’s performance (Ciarocco, Vohs, & Baumeister, 2010). Fourth, teams’ tournament performances carry high stakes, which should heighten their salience. For example, Florida Gulf Coast University’s surprisingly successful 2013 tournament performance led to a 35% increase in undergraduate applications and a 6-year, multimillion-dollar contract for coach Andy Enfield to join a higher profile program at University of Southern California (Brennan, 2014; Helfand, 2016). For the players, the tournament is a major opportunity to get noticed by NBA scouts, with potential millions in compensation at stake.

Sample

Our sample consisted of 34 seasons of college basketball, played inclusively between the 1979–1980 and 2012–2013 seasons by the 73 teams in the six major conferences during that time period (Atlantic Coast Conference [ACC], Big 12, Big East, Big 10, Pac-10, and Southeastern Conference [SEC]). This represents 2,409 possible time-lagged pairs of tournament observations (rival team prior year, focal team subsequent year). Rivalry is known to exist between teams within these conferences, driven by school similarity, repeated competition, and evenly matched competition (Kilduff et al., 2010).

We chose to collect data beginning with the 1979–1980 season as it represents the first year that the major six conferences existed in their modern form. We chose to end our data collection in the 2012–2013 season because major conference realignments occurred in the following season, disrupting rival relationships and changing the competitive dynamics within conferences. Tournament performance data were collected from the NCAA (http://fs.ncaa.org/Docs/stats/m_final4/2017/Brackets.pdf).

Measures

Rivalry

Teams’ rivals were identified from data collected by Kilduff et al. (2010) in 2005. These authors surveyed sportswriters at the student newspapers of all 73 of the universities in our sample and asked them to indicate the intensity of rivalry between their home team and each of the other teams in their conference, on a scale from 0 (not a rival) to 10 (fierce rival). Two schools responded with two or fewer surveys and were excluded from the analysis. Sportswriter responses were then validated via separate surveys of players and coaches at a subsample of universities (see Kilduff et al., 2010, p. 951). Although these rivalry data were collected at one point in time, we believe they remain valid measures of rivalry over the period of our study, as past research has shown that rivalries are quite stable over time (Kilduff et al., 2010). Further, any changes in the top rivals of teams during our period of study would work against finding support for our hypotheses.

All sportswriter ratings from each school were averaged to create a single continuous measure of the perceived intensity of rivalry that each school felt toward each other school within their conference. Then, for each school, we identified its single most intense rival.²

In analyses, we excluded from our data set any observations where the focal team and rival team were not members of the same conference in year N + 1. Although our data were drawn from a relatively stable period of time in terms of conference composition, there were still a number of instances in which teams changed conferences. Thus, there were some periods of time in which rival teams as assessed in 2005 were not regularly competing and were thus unlikely to be rivals, as repeated competition is critical to rivalry (Kilduff, 2014; Kilduff et al., 2010). For example, Florida State University joined the ACC in the 1991–1992 season. Prior to this season, Florida State University did not regularly play its primary rival as assessed in 2005, Duke. Therefore, all observations prior to the 1991–1992 season for Florida State University were excluded. Our final data set consisted of 2,086 observations.

Predictor variable: rival team’s performance in year N

Our main independent variable was the tournament performance of the focal team’s rival in the prior year (year N). If a team’s rival made the tournament, we coded that performance as a 1, and for each win a team’s rival earned during the tournament, we incremented performance by 1. Thus, this variable ranged from 0 (rival did not make the tournament) to 7 (rival won the tournament).

Because focal teams may choose rivals on the basis of similar ability, there may be a spurious correlation between focal team’s and rival team’s tournament performance. Therefore, we centered rival tournament performance in year N with respect to the rival’s mean tournament performance over our sample.

Dependent variable: focal team’s performance in year N + 1

Our dependent variable was the focal team’s tourna-ment performance in the subsequent year (year N + 1), measured on the same continuous scale (0–7) as used for the predictor variable.

Control variables

We collected control variables that may predict teams’ performance independently of their rival’s performance in the prior year (year N). We expected a team’s performance to be highly correlated year over year and wanted to control for any tendencies to perform consistently at the same level. Therefore, we controlled for the focal team’s year N win percentage net of tournament performance (collected from College Basketball Reference; www.sports-reference.com/cbb) and year N tournament performance. Because these control variables were time lagged, we centered them relative to the team’s mean season win percentage and tournament performance, respectively, over our sample.

Fixed effects

We also wanted to address any potential unobserved heterogeneity between schools that may have driven the results. For example, Tom Izzo of Michigan State is sometimes referred to as “Mr. March” for the deep runs his teams frequently make in the tournament, despite oftentimes modest expectations. Thus, we included team fixed effects in our analysis. Finally, to address any potential variability between seasons, we controlled for season fixed effects.

Results

Table 1 provides correlations and descriptive statistics for all variables. All regressions included control variables (focal team’s win percentage in year N, focal team’s tournament performance in year N) and fixed effects (team, season); regressions used heteroskedasticity- and autocorrelation-consistent (HAC) standard errors (Wooldridge, 2009).

Table 1.

NCAA Division I Men’s Basketball: Correlations and Descriptive Statistics (Study 1; N = 2,086)

Variable	M	SD	Correlations
Variable	M	SD	1	2	3	4
1. Focal team’s tournament performance in year N + 1	1.16	1.63	—
2. Rival team’s tournament performance in year N	1.77	1.98	.10**	—
3. Focal team’s tournament performance in year N	1.16	1.63	.43**	.050*	—
4. Focal team’s win percentage in year N	0.60	0.16	.46**	.030	.71**	—
5. Focal team’s seed in year N + 1	5.35	3.15	−.52**	−.030	−.30**	−.34**

Note: NCAA = National Collegiate Athletic Association.

p < .05. **p < .01.

Rival performance

In Model 1 (see Table 2), rival team’s tournament performance in year N positively predicted focal team’s tournament performance in year N + 1, b = 0.040, SE = 0.020, t(1980) = 2.03, p = .043, 95% confidence interval (CI) = [0.0013, 0.078]. These results were robust to using team random effects instead of team fixed effects (Model 2; p = .046). These results were also robust to controlling for focal team tournament seed in year N + 1 (Model 3; p = .019).³

Table 2.

NCAA Division I Men’s Basketball: Rival Team’s Tournament Performance in Year N as a Predictor of Focal Team’s Tournament Performance in Year N + 1 (Study 1)

	Model 1: fixed effects (df = 1980)		Model 2: random effects (df = 2050)		Model 3: controlling for focal team seed in N + 1 (df = 854)
Predictor	b	SE	b	SE	b	SE
Rival team’s tournament performance in year N	0.040*	0.020	0.040*	0.020	0.060*	0.025
Focal team’s tournament performance in year N	0.076*	0.034	0.076*	0.035	−0.022	0.039
Focal team’s win percentage in year N	2.52***	0.26	2.52***	0.26	0.77	0.47
Focal team’s seed in year N + 1					−0.22***	0.014

Note: All three models included year fixed effects. Models 1 and 3 contained team fixed effects, only Model 2 contained team random effects, and only Model 3 controlled for focal team’s seed in year N + 1. NCAA = National Collegiate Athletic Association.

p < .05. ***p < .001.

Rival exceptional wins

Our main model assumed that each additional postseason win by a rival has a linear additive effect on the focal team’s motivation the next year. To test whether increasingly exceptional rival performance might have an exponentially powerful effect on subsequent focal team performance, we ran a model that added the following quadratic term: year N rival team’s tournament performance squared. The quadratic term was not significant, b = −0.0053, SE = 0.0085, t(1979) = −0.63, p = .53, 95% CI = [−0.022, 0.011], suggesting that additional rival postseason wins do not exponentially increase focal team motivation.

As a further test of whether later wins matter more, we performed a linear regression analysis predicting focal team’s tournament performance in year N + 1 with seven dummy variables corresponding to each level of rival team’s tournament performance in year N (reference group = rival did not make the tournament; Table 3). We then ran a series of F tests comparing the coefficient sizes of every possible pair of rival wins (Table 3). The coefficient for rival team winning the tournament was largest; further, it was significantly greater than the coefficient for rival team making the tournament (Level 1), p = .037, and making but losing the final game (Level 6), p = .039, and marginally different from failing to make the tournament (Level 0), p = .058. The effect of a rival advancing to the fourth round was also marginally greater than when a rival did not make the tournament, p = .094. Although these results should be interpreted with caution because these coefficient comparisons have lower statistical power than our main tests, they provide some evidence that when a rival wins the tournament, it can be especially motivating.

Table 3.

NCAA Division I Men’s Basketball: Categorical Rival Team’s Tournament Performance in Year N as a Predictor of Focal Team’s Tournament Performance in Year N + 1 (Study 1; df = 1974)

Predictor	b	SE
Rival team made tournament in year N (Level 1)	−0.049_a	0.091
Rival team advanced to second round in year N (Level 2)	−0.010_a,b	0.097
Rival team advanced to third round in year N (Level 3)	0.061_a,b	0.10
Rival team advanced to fourth round in year N (Level 4)	0.22^†_a,b	0.13
Rival team advanced to fifth round in year N (Level 5)	0.19_a,b	0.17
Rival team advanced to sixth round in year N (Level 6)	−0.12_a	0.18
Rival team won tournament in year N (level 7)	0.40^†_b	0.21
Focal team’s tournament performance in year N	0.077*	0.034
Focal team’s win percentage in year N	2.50***	0.26

Note: The model contained fixed effects for year and team. For the linear regression, the reference category was rival did not make the playoffs in year N. Coefficients for rival team’s tournament performance that do not share a subscript letter were significantly different at the p < .05 level in a series of F tests. NCAA = National Collegiate Athletic Association.

†

p < .10. *p < .05. ***p < .001.

Discussion

Analyses of 34 years of college basketball data found evidence for the long shadow of rivalry. Rival team’s performance in year N predicted focal team’s performance in year N + 1. This effect was particularly pronounced when the rival won the tournament. It is also interesting to note that the motivation boost from a rival winning the championship was significantly greater than the impact of a rival making but losing the championship game; we return to this point in the General Discussion.

Study 2: Big Four U.S. Professional Sports (NBA, NFL, MLB, NHL)

Although Study 1 provided evidence for the long shadow of rivalry, the effect might have been unique to college basketball. In Study 2, we sought to replicate the findings of Study 1 in a professional context over a broader range of sports. Specifically, we examined whether rival team’s playoff performance in year N predicted focal team’s playoff performance in year N + 1 in the Big Four U.S. professional sports (NBA, NFL, MLB, NHL).

Method

Study context

In the United States, the Big Four professional sports leagues are the NBA, NFL, MLB, and NHL. Each of the Big Four leagues has 30 to 32 teams organized into conferences or divisions. Most teams earn more than $100 million in revenue per year, and in the NFL, the average is more than $250 million (Gaines, 2014). By comparison, the next largest team sports league in the United States, Major League Soccer, has just 22 teams, and the highest earning team grosses just $58 million per year (Forbes, 2016).

Each league completes each season with a postseason tournament (the playoffs) to identify its champion. Each of these playoffs is three to four rounds of head-to-head competitions between the best teams in the regular season of each league. Teams advance by winning single elimination games (NFL) or “best of” series (NBA, MLB, and NHL).

Just like the March Madness tournament, the playoffs are highly meaningful for fans and players alike. For fans, the playoffs are must-watch television, with typical viewership in the tens of millions (Gaines, 2016; Huddleston, 2017). Advertisers spend millions of dollars for 30-s ad spots in some championship games. For players, the playoffs offer both monetary rewards and reputational payoffs. Players and coaches earn performance bonuses on the basis of playoff appearances and advancement in each round, and franchise dynasties are declared and player legacies are cemented (or questioned) on the basis of playoff performance.

We expect playoff performance to be a particularly salient outcome on which rivals compare themselves for the same reasons as in Study 1: Playoff performance forms the primary measure of success for a given season, is highly visible, occurs at the end of the season, and carries high monetary and psychological stakes.

Sample

Our sample consisted of all postseasons of the Big Four sports leagues since the start of each of their modern playoff structures until the most recent playoff.⁴ We excluded from analysis any season where a playoff did not occur in year N and year N + 1 (because of labor disputes). This represents 3,292 possible time-lagged pairs of playoff observations (rival team prior year, focal team subsequent year). Playoff performance data were collected from https://www.sports-reference.com/.

Measures

Rivalry

Teams’ rivals were identified from online lists of the number-one rival for each sports team.⁵ These lists were chosen as the first web page returned by Google for the search term “top rival for each [NFL/NBA/MLB/NHL] team.” If a list mentioned more than one rival per team, we chose the rival identified as “more traditional.” If no team was mentioned as “more traditional,” we selected the first team mentioned. We excluded from analysis any observations in which the focal team or the rival team did not exist in year N or year N + 1. Our final data set contained 3,012 complete observations.

Since franchises change cities and names over time, we needed to code the rival of “extinct” teams, that is, teams that currently play under a different name or in a different city (e.g., the Seattle Supersonics). In these instances, we assigned extinct teams the rival of the contemporary franchise that, per the league, “owns” the history of the extinct team. For example, the Seattle Supersonics are extinct, but their history is owned by the Oklahoma City Thunder; therefore, we assigned the Thunder’s rival (the Dallas Mavericks) to the Supersonics.

This is a conservative measure of rivalry in these leagues because it captures rivalry at a single point in time but is applied to a team’s whole history. Although rivalries tend to be relatively stable (Kilduff et al., 2010), city changes combined with wide variance in performance over years and decades could decrease the stability of rivalries over time.

Predictor variable: rival team’s performance in year N

Our main independent variable was the playoff performance of the focal team’s rival in the prior year (year N), measured in the same fashion as in Study 1, starting with 0 (rival did not make the playoffs) and increasing in units of 1 for each round of playoffs won. This measure went up to 5 (rival won the playoffs) for the NFL, NBA, and NHL because these sports have four playoff rounds. This measure went up to 4 (won the playoffs) in MLB because MLB has three playoff rounds. We again centered rival playoff performance in year N with respect to the rival’s mean playoff performance across the sample to address the possibility of any spurious correlation driven by teams choosing rivals of similar ability.

Dependent variable: focal team’s performance in year N + 1

Our dependent variable was focal team’s playoff performance in the subsequent year (year N + 1), measured on the same scale (0–5) as was used for the predictor variable.

Control variables

As in Study 1, we controlled for focal team’s win percentage and playoff performance in year N to account for possible correlations in team performance over time. Because these control variables are time lagged, we centered them relative to the team’s mean year N win percentage and year N playoff performance across the sample.

Fixed effects

As in Study 1, we included team and season fixed effects to address potential unobserved heterogeneity between teams and seasons that may drive the results. The team fixed effects also accounted for any potential unobserved heterogeneity between sports.

Results

Table 4 includes correlations and descriptive statistics for all variables. All regressions included control variables and fixed effects and used HAC standard errors (Wooldridge, 2009).

Table 4.

Big Four: Correlations and Descriptive Statistics (Study 2; N = 3,012)

Variable	M	SD	1	2	3
1. Focal team’s playoff performance in year N + 1	0.97	1.31	—
2. Rival team’s playoff performance in year N	1.40	1.53	.070**	—
3. Focal team’s playoff performance in year N	0.97	1.31	.42**	.030	—
4. Focal team’s win percentage in year N	0.51	0.14	.40**	.020	.68**

p < .01.

Rival performance

In Model 1 (see Table 5), rival team’s playoff performance in year N positively predicted focal team’s playoff performance in year N + 1, b = 0.033, SE = 0.015, t(2780) = 2.18, p = .029, 95% CI = [0.0034, 0.063]. This result was robust to using team random effects instead of team fixed effects (Model 2; p = .035).

Table 5.

Big Four: Rival Team’s Playoff Performance in Year N as a Predictor of Focal Team’s Playoff Performance in Year N + 1 (Study 2)

	Model 1: fixed effects (df = 2780)		Model 2: random effects (df = 2901)
Predictor	b	SE	b	SE
Rival team’s playoff performance in year N	0.033*	0.015	0.033*	0.015
Focal team’s playoff performance in year N	0.15***	0.027	0.15***	0.028
Focal team’s win percentage in year N	2.09***	0.21	2.10***	0.21

Note: Both models contained year fixed effects. Only Model 1 contained team fixed effects, and only Model 2 contained team random effects.

p < .05. ***p < .001.

Rival exceptional wins

As in Study 1, we examined whether exceptional rival performance might be especially motivating. First, we tested for a quadratic relationship among our variables. The quadratic term, rival team’s year N playoff performance squared, was marginally significant, b = 0.018, SE = 0.011, t(2779) = 1.72, p = .085, 95% CI = [−0.0025, 0.039], providing some indication that when rival teams reach the later rounds of the postseason, this has an exponential effect on the subsequent postseason performance of the focal teams.

As in Study 1, we also ran a model using a set of five dummy variables for each level of rival team’s playoff performance in year N (reference group = rival did not make the playoffs) and compared the relative size of the coefficients in a series of F tests (Table 6). We excluded MLB from this analysis because it has only three playoff rounds, but results do not meaningfully change with their inclusion. These tests revealed that the coefficient for the rival winning the playoffs was significantly greater than the coefficient for the rival not making the playoffs, the rival making the playoffs, and the rival advancing one or two rounds (Levels 0, 1, 2, and 3), all ps < .05. Although these results should be interpreted with caution because of the reduced statistical power of these coefficient comparisons, they suggest that when a rival wins the playoffs in year N, it can be particularly motivating for the focal team in year N + 1.

Table 6.

Big Four: Categorical Rival Team’s Playoff Performance in Year N as a Predictor of Focal Team’s Playoff Performance in Year N + 1 (Study 2)

	Model 1: excluding MLB (df = 2348)		Model 2: all sports (df = 2776)
Predictor	b	SE	b	SE
Rival team made playoffs in year N (Level 1)	0.013_a	0.067	0.029_a	0.059
Rival team advanced to second round in year N (Level 2)	0.019_a	0.071	0.032_a	0.064
Rival team advanced to third round in year N (Level 3)	−0.016_a	0.086	0.025_a,b	0.080
Rival team advanced to fourth round in year N (Level 4)	0.13_a,b	0.11	0.086_a,b	0.096
Rival team won playoffs in year N (Level 5)	0.24*_b	0.10	0.25*_b	0.099
Focal team’s playoff performance in year N	0.16***	0.029	0.15***	0.027
Focal team’s win percentage in year N	2.02***	0.22	2.08***	0.21

Note: For the linear regression, the reference category was rival did not make the playoffs in year N. Both models contained fixed effects for year and team. Coefficients for rival team’s playoff performance that do not share a subscript letter were significantly different at the p < .05 level in a series of F tests. MLB = Major League Baseball.

p < .05. ***p < .001.

Discussion

The analyses of Study 2 revealed evidence for the long shadow of rivalry across the four major professional sports leagues in the United States. Professional sports teams performed better in the playoffs when their rival performed well in the preceding playoffs. Additional analyses suggested that a rival’s success at the highest level—winning the championship—especially motivated a focal team’s performance the following year.

General Discussion

The current research provided evidence that rivalry has a long motivational shadow. We examined two high-stakes settings—the NCAA Division I Men’s Basketball Tournament and the Big Four U.S. professional sports leagues’ playoffs—and found that a rival’s performance motivates a focal team’s performance a year later and in competition against nonrival teams. Despite substantial objective stakes, rival teams’ performance appeared to serve as critical input into teams’ motivation and performance. Furthermore, we found that a rival winning the championship was especially motivating, consistent with our opening example of the adjacent Duke and UNC titles. Although the current studies provide reliable support for our hypotheses, we acknowledge that the effect sizes are not particularly large and that many factors beyond rivalry influence postseason performance.

This research makes several theoretical contributions to the literatures on motivation and competition. First, past work has focused on the motivating effect of rivalry in head-to-head competition (e.g., Kilduff, 2014). In contrast, the present research highlights that a rivals’ performance against third parties can motivate focal actors in contests against nonrivals. In the NCAA in particular, the tournament design postpones matchups between same-conference teams, so most teams never face their rivals during the tournament. Recent research suggests that merely thinking about a rival can produce rivalry’s consequences (Kilduff & Galinsky, 2017). Overall, the evidence suggests that rivalry can escape the bounds of head-to-head competition and cast its shadow into situations that do not include direct competition with one’s rival.

We believe that rival performance becomes a specific, salient goal for focal actors that fuels their performance. Although we were unable to measure goal setting with the current data, we consider the conceptual link between rivalry and goal setting to be an important theoretical contribution given the powerful impact of goals (e.g., Locke & Latham, 1990).

Also consistent with the link between rivalry and goal setting, our results show that rivalry can affect long-term behavior, extending existing research on rivalry’s immediate consequences. At the close of the postseason, focal teams know that their first chance to draw even with a successful rival is 12 months away. Despite this time lag, rivalry can fan the motivational flames to generate greater effort 1 year later.

The present research also adds to the growing evidence of the power of rivalry to impact group-level performance and outcomes (Kilduff et al., 2010; Kilduff et al., 2016). Past findings on how rivalry impacts competitive success have primarily been at the individual level (but see defensive efficiency in Kilduff et al., 2010). We extend this research to show that rivalry affects group-level performance as well. This extension to group-level performance is consistent with past literature that finds that interactions are more competitive at the group than individual level (Insko et al., 1994; Meier & Hinsz, 2004; Rabbie, 1998).

Finally, our findings suggest that the broader competitive success of rivals is intertwined, their fates linked. High performance by a rival inspires high performance by a focal actor. But rivalry also has the potential to demotivate; when one’s rival performs poorly, a focal actor may lower his or her own aspirations. This may be especially true for poor rival performance in the most salient or high-stakes of contests. Indeed, in the college basketball data, the boost of a rival winning the championship was significantly greater than a rival making but losing the championship game. Although this finding should be interpreted with caution given that the effect was not replicated in the professional context, it is consistent with schadenfreude research, which has found that watching a rival lose in the World Cup after your team had been eliminated was especially satisfying (Leach, Spears, Branscombe, & Doosje, 2003). Because of the high visibility and stakes of the championship game, watching a rival lose this game may be so satisfying that it fosters complacency, effectively eliminating the motivational boost that could have resulted from the rival’s exceptional performance.

This highlights a critical distinction between competitors and rivals: Competitors are actors with negative interdependence between their outcomes (Deutsch, 1949), whereas rivals, outside of head-to-head competition, may actually experience positive interdependence. Although rivals are often thought of as the fiercest of competitors, their relationship and psychological connection link them together outside of head-to-head contests. Their boats appear to rise and fall together over time.

Limitations and future directions

As with all archival research, we are limited in our ability to establish causality. The time-lagged nature of our effect, the robustness of our findings to a variety of controls and econometric tools, and replication across amateur and professional contexts and across sports give us some confidence that a rival’s performance causally impacts a focal team’s performance. However, future experimental research could more definitively establish causality. Further, our data did not allow us to test our proposed mechanism of goal setting. Future work should directly examine the link between rivalry and goals.

Conclusion

Archival analyses of the NCAA Division I Men’s Basketball Tournament and the Big Four U.S. professional sports leagues’ playoffs demonstrated that rivalry motivates performance in non-head-to-head competition 1 year later. Across a variety of contexts, a rival team’s performance in the prior postseason predicted a focal team’s performance in the following postseason. Rivalry truly has a long shadow.

Footnotes

Action Editor

Bill von Hippel served as action editor for this article.

Author Contributions

B. E. Pike and A. D. Galinsky developed the study concept. All authors contributed to the study design. Testing and data collection were performed by B. E. Pike and G. J. Kilduff. B. E. Pike analyzed and interpreted the data under the supervision of A. D. Galinsky and G. J. Kilduff. B. E. Pike drafted the manuscript, and all authors provided critical revisions. All authors approved the final version of the manuscript for submission.

Declaration of Conflicting Interests

The author(s) declared that there were no conflicts of interest with respect to the authorship or the publication of this article.

Open Practices

All data have been made publicly available via the Open Science Framework and can be accessed at https://osf.io/5vy69/?view_only=23a8c0762ef64333b1eb605010d68934. The complete Open Practices Disclosure for this article can be found at http://journals.sagepub.com/doi/suppl/10.1177/0956797617744796. This article has received the badge for Open Data. More information about the Open Practices badges can be found at .

Notes

References

Brennan

(2014, January 8). Florida Gulf Coast’s enrollment spikes. ESPN. Retrieved from http://www.espn.com/blog/collegebasketballnation/post/_/id/92235/fgcus-enrollment-did-exactly-what-youd-expect

Ciarocco

N. J.

Vohs

K. D.

Baumeister

R. F.

(2010). Some good news about rumination: Task-focused thinking after failure facilitates performance improvement. Journal of Social & Clinical Psychology, 29, 1057–1073.

Converse

B. A.

Reinhard

D. A.

(2016). On rivalry and goal pursuit: Shared competitive history, legacy concerns, and strategy selection. Journal of Personality and Social Psychology, 110, 191–213.

Deutsch

(1949). An experimental study of the effects of co-operation and competition upon group process. Human Relations, 2, 199–231.

Forbes. (2016, September 7). Forbes releases 2016 MLS team valuations. Forbes. Retrieved from https://www.forbes.com/sites/forbespr/2016/09/07/forbes-releases-2016-mls-team-valuations/#6c3a47314f85

Gaines

(2014, May 14). CHART: NFL and MLB teams top premier league clubs when it comes to making money. Business Insider. Retrieved from http://www.businessinsider.com/chart-nfl-mlb-premier-league-revenue-2014-5

Gaines

(2016, November 2). World Series TV ratings are soaring and Game 7 is expected to be a monster. Business Insider. Retrieved from http://www.businessinsider.com/world-series-tv-ratings-history-2016-11

Galinsky

Schweitzer

(2015). Friend & foe: When to cooperate, when to compete, and how to succeed at both. New York, NY: Crown Business.

Harackiewicz

J. M.

Barron

K. E.

Tauer

J. M.

Elliot

A. J.

(2002). Predicting success in college: A longitudinal study of achievement goals and ability measures as predictors of interest and performance from freshman year through graduation. Journal of Educational Psychology, 94, 562–575.

10.

Heath

Larrick

R. P.

(1999). Goals as reference points. Cognitive Psychology, 38, 79–109.

11.

Heitner

(2017, January 31). $4.5 billion will be wagered illegally on Super Bowl 51. Forbes. Retrieved from https://www.forbes.com/sites/darrenheitner/2017/01/31/4-5-billion-will-be-wagered-illegally-on-super-bowl-51/#55338f2c217f

12.

Helfand

(2016, March 18). Andy Enfield and USC agree to a two-year contract extension. Los Angeles Times. Retrieved from http://www.latimes.com/sports/usc/la-sp-usc-basketball-andy-enfield-contract-extension-20160328-story.html

13.

Huddleston

Jr. (2017, February 6). Here’s how many people watched the Super Bowl. Fortune. Retrieved from http://fortune.com/2017/02/06/super-bowl-111-million-viewers/

14.

Insko

C. A.

Schopler

Graetz

K. A.

Drigotas

S. M.

Currey

D. P.

Smith

. . . Bornstein

(1994). Interindividual-intergroup discontinuity in the prisoner’s dilemma game. Journal of Conflict Resolution, 38, 87–116.

15.

Kahneman

Fredrickson

B. L.

Schreiber

C. A.

Redelmeier

D. A.

(1993). When more pain is preferred to less: Adding a better end. Psychological Science, 4, 401–405.

16.

Kilduff

G. J.

(2014). Driven to win: Rivalry, motivation, and performance. Social Psychological & Personality Science, 5, 944–952.

17.

Kilduff

G. J.

Elfenbein

H. A.

Staw

B. M.

(2010). The psychology of rivalry: A relationally dependent analysis of competition. Academy of Management Journal, 53, 943–969.

18.

Kilduff

G. J.

Galinsky

A. D.

(2017). The spark that ignites: Mere exposure to rivals increases Machiavellianism and unethical behavior. Journal of Experimental Social Psychology, 69, 156–162.

19.

Kilduff

G. J.

Galinsky

A. D.

Gallo

Reade

J. J.

(2016). Whatever it takes to win: Rivalry increases unethical behavior. Academy of Management Journal, 59, 1508–1534.

20.

Leach

C. W.

Spears

Branscombe

N. R.

Doosje

(2003). Malicious pleasure: Schadenfreude at the suffering of another group. Journal of Personality and Social Psychology, 84, 932–943.

21.

Locke

E. A.

Latham

G. P.

(1990). A theory of goal setting & task performance. Upper Saddle River, NJ: Prentice Hall.

22.

Meier

B. P.

Hinsz

V. B.

(2004). A comparison of human aggression committed by groups and individuals: An interindividual–intergroup discontinuity. Journal of Experimental Social Psychology, 40, 551–559.

23.

NCAA.com. (2017a, April 7). 2017 Final Four viewership, attendance by the numbers. NCAA.com. Retrieved from http://www.ncaa.com/news/basketball-men/article/2017-04-07/2017-final-four-viewership-attendance-numbers

24.

NCAA.com. (2017b, March 27). 2017 NCAA tournament scores increases in television viewership through Elite 8, along with all-time records across digital and social platforms. NCAA.com. Retrieved from http://www.ncaa.com/news/basketball-men/article/2017-03-27/2017-ncaa-tournament-scores-increases-television-viewership

25.

Neely

(2012, September 17). Greatest Lakers rivalry: Larry Bird vs. Magic Johnson. Rant Sports. Retrieved from http://www.rantsports.com/nba/2012/09/17/greatest-lakers-rivalry-larry-bird-vs-magic-johnson/

26.

Novak

(2013, July 28). Ranking the 25 best rivalries in college basketball. Bleacher Report. Retrieved from http://bleacherreport.com/articles/1719358-ranking-the-25-best-rivalries-in-college-basketball

27.

Paine

(2017, March 3). UNC-Duke really is college basketball’s best rivalry. FiveThirtyEight. Retrieved from https://fivethirtyeight.com/features/unc-duke-really-is-college-basketballs-best-rivalry/

28.

Purdum

(2017, March 14). 70 million brackets, $10.4 billion in bets expected for March Madness. ESPN. Retrieved from http://www.espn.com/chalk/story/_/id/18901384/70-million-brackets-104-billion-bet-ncaa-tournament

29.

Rabbie

J. M.

(1998). Is there a discontinuity or a reciprocity effect in cooperation and competition between individuals and groups? European Journal of Social Psychology, 28, 483–507.

30.

Redelmeier

D. A.

Kahneman

(1996). Patients’ memories of painful medical treatments: Real-time and retrospective evaluations of two minimally invasive procedures. Pain, 66, 3–8.

31.

Rosenberg

(2017, April 4). UNC overcomes ghosts, pain, Gonzaga to reach its national title ceiling the Carolina Way. Sports Illustrated. Retrieved from https://www.si.com/college-basketball/2017/04/04/unc-national-champions-north-carolina-gonzaga-hicks-berry-roy-williams

32.

Kilduff

G. J.

Ordonez

Schweitzer

M. E.

(2017). Going for it on fourth down: Rivalry increases risk-taking, physiological arousal, and promotion focus. Academy of Management Journal. Advance online publication. doi:10.5465/amj.2016.0850

33.

Wooldridge

J. M.

(2009). Introductory econometrics: A modern approach (4th ed.). Mason, OH: Cengage Learning.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.78 MB