Awards and status

In most cases, awards are markers of excellence and are celebrated and announced in a public fashion [3]. Winning an award sets the recipient apart from their peers and can have positive implications for their careers. Often, winning an award results in a special form of social distinction that can spread and grow over time [11]. As long as prospective recipients value the award and have a desire to obtain it, recipients can enjoy status and reputational and recognitional benefits as a result of being connected to the award [1,11]. This recognition may be quite enduring, impactful, and can even grow as more people become aware of and recognize a winner’s distinction.

Recognition that is conferred by the eminence of the award has also been referred to as external status. External status can be defined as how broad one’s visibility is outside of their own organization [12]. Employees attain external status as the result of attributes independent of the level or visibility of their current task performance [12]. A sub-type of external status, affiliation-based external status, does not stem from a demonstrated quality or level of performance, but rather refers to status that is inherited as a result of a close personal connection with and/or sponsorship by an elite individual or institution [12,13]. Affiliation-based external status can be considered the reputational benefits and favorable attributions one is ascribed on the basis of their associations (e.g., the prestige of the university they attended, the notoriety of an award they won). The exclusivity and reputation of the affiliation is thought to be correlated with the value gained by the affiliation [12,13]. Namely, affiliations perceived as prestigious will result in more status than lesser-known or easier to obtain affiliations.

The benefits of having affiliation-based external status are numerous. Affiliation-based external status serves as a signal or message to outsiders, regardless if they are a subject matter expert or not, that the award winner is admirable [3]. The heightened perception of the individual, in turn, influences perceptions of the quality of the work the individual produces [14]. Status creates the assumption of higher quality and may implicitly result in greater attention to the work of those with status [4]. Thus, award winners can receive more favourable judgments as a result of their affiliation-based external status and this can result in more positive career outcomes through increased attention and access to resources.

Opportunity, productivity and the Matthew Effect

Being the recipient of an award and benefitting from its positive effects is consistent with the Matthew Effect [4,5,13]. The Matthew Effect proposes that receiving an award can potentially create inequalities between two individuals of relatively similar ability by way of its conferred status over time. When award-winning scientists may begin to receive peer recognition, the process of cumulative advantage is thought to begin. In a process similar to compound interest, an initial comparative advantage can compound to yield larger, successive increments of advantage such that the gaps between those who have resources and those who do not widen over time [15,16]. Increased recognition may lead to successively greater opportunities for advancing work, which, in turn, results in greater productivity and the rewards that go with increased productivity [5]. Thus, an initial award win may result in winners, when compared to a finalist for that same award, being more productive and earning more resources over the course of their career as a result of a potentially small initial advantage.

Numerous analytical approaches have been utilized to study the Matthew Effect and its core tenets. Often, researchers contrast award winners’ post-award productivity with a control group of non-award-winning researchers of commensurate abilities [8,9,14]. Prior empirical work has applied matching techniques and have matched either work authored by an award winner [17] or award-winning authors themselves [10] with controls and compared post-award productivity. Researchers have also utilized within-person approaches: some have contrasted individual author’s pre and post award career outcomes [18] and others have utilized formal statistical tests (e.g., Gilbrat’s law; the Kolmogorov–Smirnov test) to analyze citation patterns of authors for post-award deviations [19,20]. Other work has investigated how awards impact the attainment of more objective career success measures such as promotions or salary increases [2,7]. Interestingly, previous findings on how receiving an award influences future productivity have been mixed [6–10] and little is known about how components of the Matthew Effect, specifically opportunity, operate.

More specifically, there are findings that both support and contradict the Matthew Effect being initiated by winning an award. Some research supports the idea that winning an award facilitates subsequent productivity (e.g., publications, future awards) [6,7,18], but there is also evidence to suggest award winners are actually less productive post-win [8–10,17]. Also, despite the importance of opportunities in the Matthew Effect, it is not clear if winning an award affects subsequent opportunities and in turn, how opportunities influence productivity and future rewards. Studies evaluating the Matthew Effect frequently compare measures of post award productivity and often do not gather nor evaluate measures of opportunities pre and/or post award [9,10]. Together, previous empirical research suggests that the effect of awards on recipients’ career outcomes may not be straightforward, and some unknowns still exist.

The current study

The effect of awards on career outcomes is likely a dynamic process that unfolds over a long period of time and involves an interaction between a number of psychological, perceptual, and decision-making processes both within and outside of the recipient. One of the main challenges of interpreting empirical results on the effect of awards on outcomes is the confounding nature of award winners’ endogenous characteristics with the receipt of an award. That is, trying to answer the questions of if, and the extent to which, affiliation-based external status from awards and recognitions uniquely contributes to career outcomes is inherently difficult because it is generally recognized that receiving an award is the result of some type of demonstrated excellence. Indeed, most of the research on awards examines awards that are achievement-based (e.g., Fields Medal, Nobel Prize). As a result, winners are presumed to be more deserving of an award than non-winners, via their better performance, achievements, qualifications, etc. Therefore, any subsequent differences in career outcomes between an award winner and non-winners may simply be considered a reflection of individual differences that helped lead to the receipt of the award.

Disentangling the effect of winners’ endogenous characteristics and affiliation-based external status on career outcomes is a noted challenge within the Matthew Effect literature. For example, Bol and colleagues [6] purport that research on the Matthew Effect has been hampered by the possibility that observed cumulative differences in achievement may in fact be the gradual manifestation of interpersonal differences in talent or productivity. Azoulay and colleagues [14] suggest that an individual’s quality and status are typically tightly intertwined. In doing so, they question if status affects outcomes or if status is simply a by-product of an individual’s positive endogenous qualities.

Pinpointing and examining the onset of affiliation-based external status can provide a unique opportunity to assess the Matthew Effect and its particular impact on career outcomes. Recall, affiliation-based external status occurs when an individual enjoys broad attention as a result of a close affiliation with an elite person or institution and is independent of the level or visibility of the recipient’s task performance [4,12]. Whereas most of the literature on the Matthew Effect has examined the impact of status achieved by winning performance-based awards, affiliation-based external status is a specific type of status that can be widely recognized, but can be achieved without demonstrating a high level of performance or productivity in a particular domain. As a result, being able to identify the onset of affiliation-based external status can provide a scenario where one can compare post-status acquisition outcomes against those who did not receive the status, while not conflating outcomes with other endogenous characteristics.

A championship in a team sport is a team award and therefore is not specifically a marker of individual team members’ task performance; yet, the team award provides affiliation-based external status to all individual winners: all players who win are part of a small, elite, recognizable group who have won the championship. Most professional sports leagues (e.g., the National Basketball Association, the National Hockey league) actively work to prevent all of the top players from ending up on the same team by instituting salary management practices (e.g., wage caps) [21]. Whereas differences in career outcomes after winning an individual award are often conflated with ability prior to receiving the award, the structure of professional sports necessitates winning teams having players of differing abilities. Accordingly, both winning teams and non-winning teams consist of a large number of players with variable skill and productivity levels. This allows for naturally occurring control groups, where one can find players from winning and non-winning teams with roughly equivalent endogenous characteristics. Thus, award winners can be matched with non-award winners of commensurate ability and comparisons can be made post-award win to determine the unique impact of the award on career outcomes without conflating ability. Considering league championships are awards that are not allocated solely based on individual productivity, but still transmit a substantial amount of status, they may be utilized as a marker of external status that is irrespective of endogenous characteristics prior to winning the award.

Examining the Matthew Effect in the context of professional sports has other advantages. Sports data provide a number of objective measures to study the Matthew Effect: awards, opportunities, and productivity. These data also facilitate the examination of phenomena over time as individual player statistics are recorded and logged across careers. This allows researchers to use behavioral data from real-world contexts and makes it possible to create relatively large samples [22]. The meticulous tracking and recording of statistics over players’ careers and objective measures of affiliation-based external status, individual productivity, and individual opportunity makes this type of data an ideal fit for determining how affiliation-based external status uniquely affects career outcomes of individuals.

The current research provides a mechanistic investigation of Matthew Effect by examining if winning an award that confers affiliation-based external status is associated with differences in career outcomes between those who receive the award and those who do not. In doing so, we address two important limitations noted from previous work. (a) By examining an award that is not conflated with individual’s positive endogenous characteristics it will be possible to address concerns raised when interpreting previous research. (b) Furthermore, the current research will address limitations of previous work by formally operationalizing and investigating opportunity as a mechanism of the effect of awards on career outcomes. Using professional sport data, we examine if being designated as “a winner” (by way of attaining a league championship) is associated with better career outcomes (i.e., opportunities, productivity) irrespective of other individual differences that are known to predict success. Specifically, this work seeks to answer:

1. Do individuals with affiliation-based external status have more opportunities across their careers than individuals of equal ability who do not have the same affiliation-based external status?
2. a) Are those with affiliation-based external status more productive across their careers than individuals of equal ability that do not have the same affiliation-based external status? b) Does opportunity partially explain differences in career productivity between those with affiliation-based external status and those without it?

To answer these questions, several analytic strategies were applied to two different datasets. In one dataset, award winners are compared to different groups of non-winners. First, award winners are compared to a group of non-winners on career opportunity and productivity statistics. Next, career outcomes between award winners and award finalists are compared. Then, a matching technique is used to compare winners with peers of similar tenure and ability at the time of the winner’s win. Subsequently, we examine whether opportunity increases are a mechanism for productivity gains using mediational analyses. Lastly, two supplementary analyses with additional data are conducted to further investigate the role of opportunity in the Matthew Effect.

About the data

To investigate these questions, data from the National Hockey League (NHL) was used. The league championship trophy of the NHL is called the Stanley Cup. Individual player statistics from National Hockey League (NHL) players drafted in the years 1996–2002 were gathered from a publicly available, online database [23]. First, a list of players selected in the 1996–2002 NHL Entry Drafts was compiled. Next, individual players were searched on Hockey Reference. Players yearly statistics and career statistics were recorded. Specifically, we recorded player’s: draft year, name, if they have ever been a Stanley Cup winner or finalist, if they’ve ever received an All-Star team award, their career length (in seasons), games played, points, and total ice time (see Table 1 for descriptive statistics). Lastly, players were then compared to exclusion criteria (noted below) and deemed eligible or not for our work.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0290954.t001

We had some exclusion criteria for our dataset. Goaltenders were excluded as they have different performance statistics than forwards and defenseman. All of the data utilized in this work pertains to regular season performance and opportunity statistics only. The NHL is structured such that all of the teams in the league play an 82-game regular season. The top performing 16 teams in the regular season advance to the play-offs. The bottom 16 teams in the league are eliminated from contention. As such, players on teams that make the play-offs have the potential to play substantially more games (ranging from 4 to 28 games, per season) and to accrue more points and ice time than players on teams that do not make the playoffs. Accordingly, this work excludes play-off statistics from consideration. Additionally, in order for a player to be recognized as a Stanley Cup winner, a player on the Stanley Cup winning team must have played at least 41 games in a single season. Therefore, players drafted within the 1996–2002 drafts who did not play in at least 41 games in a single season were excluded from the dataset as they were not eligible to have their name on the Stanley Cup. Of the 1896 players drafted in 1996–2002, 427 players were deemed eligible. 74 of those players won the Stanley Cup (17%).

Measures

Comparing winners against non-winners and finalists

Our research questions are specific to affiliation-based external status’s relation with opportunities and productivity. To examine these relations, we utilize several analyses that progress from simple to more complex in an attempt to address different interpretations and rule out possible confounds. We utilize a draft cohort strategy for our initial analyses. Players from the same draft years are roughly the same age and therefore have the same potential for opportunities and productivity and will also play against similar opponents throughout their careers. The first analysis is simple, in which we compared the career level data of Stanley Cup winners to career level data of players in the same cohorts who did not win the Stanley Cup. We used two different comparison groups. Specifically, winners were compared to non-winners (every player from the selected cohorts that never won a Stanley Cup in their career) and then compared to winners to finalists (players that reached the Stanley Cup finals at one point in their career, but never won). Finalists can be considered a closer comparison group than non-winners as they closely approximate winners (i.e., finalists that came as close as possible to winning).

Winners were compared to non-winners and finalists across two opportunity statistics (total ice time and games played) and a productivity measure (points). All comparisons were done using a series of paired t-tests. Cohen’s d, an effect size that reveals the standardized difference between two means, and corresponding confidence intervals (CI), parameters that measure the precision of the effect size, were also calculated. A commonly used interpretation is to refer to effect sizes as small (d = 0.2), medium (d = 0.5), and large (d = 0.8) based on benchmarks suggested by Cohen [26]. Results are displayed below (Tables 2 and 3).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Results of t-test and descriptive statistics comparing winners to non-winners.

https://doi.org/10.1371/journal.pone.0290954.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Results of t-test and descriptive statistics comparing winners to finalists.

https://doi.org/10.1371/journal.pone.0290954.t003

Overall, these results indicate that when comparing the career totals of winners to non-winners and finalists, winners consistently received more opportunities (i.e., played more games, had more time on ice), but were not always more productive. For opportunity measures, all analyses comparing winners with non-winners and winners with finalists reveal statistically significant and medium to large effect sizes. Winners were found to be more productive (medium to large effect, d > .5) in their careers compared to non-winners, but did not have more productive careers than finalists.

Do teams with the best players win?

To help address the perennial issues of causality in understanding the relation between awards and outcomes, additional analyses were conducted with a modified dataset, where the top performing players were removed. Every year, the Professional Hockey Writer’s Association votes for the best 2 performers over the season at each position. Thus, the 10 best players receive a performance-based award at the end of the season and are classified as All-Stars. In this analysis, anyone who had won an All-Star award (e.g., were one of the 10 best players in the league at any point during their career) was removed from the dataset. A total of 15 players were an All-Star at some point in their career and were subsequently removed from the dataset (N = 412). As a result, these analyses examined if affiliation-based external status can generate differences in career outcomes of middle to low performers. Because the current dataset excludes players that have been recognized as top performers (i.e., award winners due to their productivity) and only includes players who would not normally receive any awards for their performance, it offers a unique opportunity to see if affiliation-based status bequeaths benefits irrespective of the endogenous characteristics of award winners. As in the above analyses, opportunity and performance measures between Stanley Cup winners and non-winners and between winners and finalists were compared using a series of paired t-tests. Cohen’s d and corresponding confidence intervals were also calculated. Results of the analyses are displayed below (Tables 4 and 5).

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Results of t-test and descriptive statistics comparing winners to non-winners, excluding NHL all-stars.

https://doi.org/10.1371/journal.pone.0290954.t004

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Results of t-test and descriptive statistics comparing winners to finalists, excluding NHL all-stars.

https://doi.org/10.1371/journal.pone.0290954.t005

Together, these results mirror the results of the analyses that were conducted with the top performing players included. Winners received significantly more opportunities than non-winners and finalists and were significantly more productive than non-winners, but not more productive than finalists. Overall, our findings remain consistent regardless of the whether the top performers were included in the comparison or not.

Do differences in outcomes arise after the win?

By finding opportunity and some productivity differences across careers between those with an award and those without, the above results, appear to suggest some support for Merton’s Matthew Effect. However, a comparison of total career outcomes leaves the results vulnerable to the possibility that differences in productivity and opportunity are not the result of increased affiliation-based external status. That is, the differences in career totals could have been generated at any point in a player’s career and the extent to which affiliation-based external status specifically impacted career outcomes is not clear. To address this, career outcomes need to be analyzed after players receive affiliation-based externals status (i.e., wins the Stanley Cup). Although it is clear when status is acquired by individuals, players win the Stanley Cup at different points in their careers (i.e., some players win early in their careers, others win later). This makes it difficult to have a traditional pre post design with a control group as there is not a standardized point in which players either win or become finalists.

To address this challenge, we utilized a matching technique. In these analyses, award winners were matched at the time they won the award with a non-winner of commensurate ability and experience. Matching controls for some of or all of the confounding influence of pre-treatment control variables in observational data by creating a sample that did not receive the treatment (i.e., did not win the Stanley Cup) that is comparable on all observed levels to the sample that did receive the treatment (i.e., those that won the Stanley Cup). Then, outcomes after the players were matched can be compared to determine if affiliation-based external status generates any differences in career outcomes.

This work utilizes coarsened exact matching (CEM) methods to match Stanley Cup winners. CEM is a method for improving the estimation of causal effects by reducing imbalances in covariates between treated and control groups [27]. CEM is a monotonic imbalance-reducing matching method, which means that the variables the user wants to balance between the treated and the control groups is chosen by the user prior to completing any analyses [27]. The idea of CEM is to temporarily coarsen each variable into substantively meaningful groups, exact match on these coarsened data, and then retain only the original (un-coarsened) values of the matched data [27]. Users complete analyses on the un-coarsened data.

The first task was to choose a set of covariates to guarantee balance between the treatment group (i.e., players that won the Stanley Cup) and the control group (i.e., players that did not win the Stanley Cup). To match Stanley Cup winners with an appropriate comparison, players were matched on the number of years they have played in the NHL, their position, and their skill index. A player’s skill index is simply the number of games they’ve played plus their points. A skill index was calculated for each season players played, and a cumulative skill index was kept for each player. The winner’s cumulative skill index for the year they won the Stanley Cup was recorded and then they were matched with a player in the same season of their career, who played the same position, and had a similar skill index. For example, let’s say Player X is a forward who won the Stanley Cup in his second season. At the time of his second season, he had 50 points and 140 games played, equaling a cumulative skill index of 190. Player X would then be matched with a player that has not won the Stanley Cup, that is also a forward, in his second season that has a skill index of 190.

Exact matching was used where possible. Due to the number of player positions and the number of years in the league, exact matching could be used for those categories. Skill indexes were more variable and as such, we had to create predetermined cut-offs for matching. The predetermined cut-off for bins was 50 points on the skill index. Binning for variables should be customized according to the distribution of the variable and the bins for the variable should be created as to ensure the closest matches possible without losing a large proportion of the sample [28]. All winners, except for one who was excluded from the dataset for not having a comparable match, received a match. The mean skill index for winners was 492 and the mean skill index for non-winners was 491. Of the 73 pairs, 18 were perfect matches, 46 matches were within three points of each other, 56 matches were within five points, 64 matches were within ten points, and the remaining nine were within 43 points. From the matched players, two groups were created: one with players who have won the Stanley Cup and the other consisting of those who had not won the Stanley Cup. Four t-tests were conducted to ensure that there are not any significant differences between winners and non-winners on any of the indexes used in the matching process (Table 6). None of the tests were significant, lending support for our matching method.

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. Results of t-tests and descriptive statistics analyzing matching method.

https://doi.org/10.1371/journal.pone.0290954.t006

The effects of winning the Stanley Cup (between subject: winner vs. non-winner) and Time (within subject: pre-treatment vs. post-treatment) on Opportunities (measured by number of Games Played and Total Ice Time) were examined using two between versus within 2x2 ANOVAs. The effects of winning the Stanley Cup (between subject: winner vs. non-winner) and Time (within subject: pre-treatment vs. post-treatment) on productivity (measured by Points) was also examined using another between versus within 2x2 ANOVA. Descriptive statistics are reported in Tables 7–9.

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. Means and standard deviations for GP as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.t007

Download:

• PNG
  larger image
• TIFF
  original image

Table 8. Means and standard deviations for TOI as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.t008

Download:

• PNG
  larger image
• TIFF
  original image

Table 9. Means and standard deviations for Pts as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.t009

First, the main effect of winning the Stanley Cup on games played was examined. The main effect of winning the Stanley Cup was significant, F (2, 288) = 14.29, p = .000, partial η² = .05, 90% CI [.02, .09]. The marginal mean for Stanley Cup winners was 386.24 and the marginal mean for non-winners was 303.67, revealing that Stanley Cup winners played significantly more games in their careers than non-winners. This result is consistent with previous analyses indicating that winners have more games played over their careers than non-winners.

The main effect of time (pre/post treatment) was also statistically significant, F(2, 288) = 4.57, p = .033, partial η² = .02, 90% CI[.00, .05]. Examining the main effect of time (pre/post treatment) reveals the pre-treatment marginal mean was 350.86 and the marginal mean post-treatment was 339.05, suggesting players played more games pre-treatment than post treatment. The interaction between winning the Stanley Cup and time was also significant, F(1, 288) = 6.95, p = .009, partial η² = .02, 90% CI[.00, .06], (Fig 1 and Table 10). The significant interaction in Fig 1 reveals that there was a change in the number of games played at time two (i.e., after winning the Stanley Cup), such that those who won the Stanley Cup (M = 420.49) played more games than those who had not won the Stanley Cup (M = 269.62) post-treatment. This result is consistent with the expectation that winning results in more opportunities for winner compared to non-winners.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Group means for the interaction of games played as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.g001

Download:

• PNG
  larger image
• TIFF
  original image

Table 10. Fixed-Effects ANOVA results using games played as the criterion.

https://doi.org/10.1371/journal.pone.0290954.t010

The effects of winning the Stanley Cup on total ice time was then examined. The main effect of winning the Stanley Cup was significant, F (2, 290) = 9.46, p = .002, partial η² = .02 90% CI [.01, .07]. The marginal means of winning the Stanley Cup (6675.36) was greater than the marginal mean for non-winners (5193.97), indicating winners play significantly more minutes in their career overall. The main effect of time (pre/post treatment) was not significant, F(2, 288) = 2.39, p = .123, partial η² = .01, 90% CI[.00, .03]. The marginal means pre-treatment (6014.22) appear to be greater than the marginal mean post-treatment (5855.11), indicating that players did not play different amounts before and after the treatment. The interaction between winning the Stanley Cup and time was not significant, F(1, 288) = 3.70, p = .055, partial η² = .01, 90% CI[.00, .04], (Fig 2 and Table 11). Although the pattern of the interaction is consistent with winners playing more minutes after the award (i.e., treatment) than non-winners, and p = .055, the differences post treatment were not statistically significant at p < .05.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Group means for the interaction of total ice time as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.g002

Download:

• PNG
  larger image
• TIFF
  original image

Table 11. Fixed-effects ANOVA results using TOI as the criterion.

https://doi.org/10.1371/journal.pone.0290954.t011

Next, the effects of winning the Stanley Cup on productivity, as measured by points, was examined. The main effect of winning the Stanley Cup was significant, F (2, 290) = 7.35, p = .007, partial η² = .02 90% CI[.00, .06]. The marginal means of winning the Stanley Cup (157.55) was greater than the marginal mean for non-winners (121.17), indicating winners had significantly more points in their careers than non-winners. The main effect of time was not significant, F(2, 288) = 2.13, p = .146, partial η² = .01, 90% CI[.00, .03]. The marginal mean pre-treatment (140.59) is slightly greater than the marginal mean post-treatment (138.13), but the difference is not statistically significant. The interaction between winning the Stanley Cup and points was not significant at p < .05, F(1, 288) = 2.61, p = .054, partial η² = .01, 90% CI[.00, .04], (Fig 3 and Table 12), although the pattern of the interaction is consistent with winners acquiring more points after the treatment than non-winners.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Group means for the interaction of points as a function of a 2 (Won Cup) X 2 (Time) design.

https://doi.org/10.1371/journal.pone.0290954.g003

Download:

• PNG
  larger image
• TIFF
  original image

Table 12. Fixed-effects ANOVA results using points as the criterion.

https://doi.org/10.1371/journal.pone.0290954.t012

Together, these results suggest that when matching winners with an equivalent player with respect to tenure and ability at the time of the win, winning the Stanley Cup results in an increased number of games played post win. Although the interactions for time on ice and points were in the expected pattern, the interactions were not statistically significant at p < .05 (p = .055 and p = .054, respectively).

Is there an indirect effect on productivity through opportunities?

Although the above results demonstrated that there was not a statistically significant direct effect of status on productivity following the onset of status, the Matthew Effect posits that awards can generate status that results in more opportunities, which then lead to resource gains [4]. Thus, it is possible that opportunities may act as a mechanism for a possible indirect relation between status and productivity. In the next analysis, we examine if status has an indirect effect on productivity through opportunity. To do so, a mediation analysis was conducted to examine the possible impact of opportunity in facilitating resource gain. Using the matched paired dataset, the mediation analysis examined whether being a Stanley Cup winner (X; compared to matched non-winners at time of the win) affects the number of points (Y) through the number of games a player has played (M). In this scenario, c is the zero-order relation between being a Stanley Cup winner and points (Y). The indirect effect is the effect of being a Stanley Cup winner on points that is mediated through games played (the product of the a and b paths, a*b). To test the statistical significance of the indirect effect, the mediation package was utilized in R. Results of these analyses are depicted in Fig 4. As Fig 4 illustrates, the regression coefficient between being a Stanley Cup winner and points and the regression coefficient between games played and points were significant. The model shows a significant indirect effect of games played on the relationship between winning the Stanley Cup and points (90.36 [39.13, 143.46], p < .001) (Table 13). The Average Causal Mediation Effect (ACME), or indirect effect, represents the effect of the mediator (i.e., games played) alone on productivity and is significant. The average direct effect (ADE) describes the direct effect of being a Stanley Cup winner on points and is insignificant. The total effect (ACME + ADE) describes the total effect of the treatment on the outcome and is also significant. Prop. mediated describes the proportion of the effect of being a Stanley Cup winner on points that goes through games played. It is calculated by dividing the ACME by the total effect and is also significant. Thus, the indirect, or mediation effect, of games played was statistically significant.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Unstandardized coefficients for the relationship between being a Stanley Cup winner and points as mediated by games played.

https://doi.org/10.1371/journal.pone.0290954.g004

Download:

• PNG
  larger image
• TIFF
  original image

Table 13. Causal mediation analysis for the relationship between being a stanley cup winner and points as mediated by games played.

https://doi.org/10.1371/journal.pone.0290954.t013

Upsilon() was then utilized to determine the effect size of the indirect effect. The effect size is a method for quantifying explained variance in mediation analysis, and is interpreted as the variance in Y accounted for jointly by M and X that corrects for spurious correlation induced by the ordering of variables [29]. In the current analyses, is the estimated variance in points explained by winning the Stanley Cup through games played and was 0.063 with a 95% CI [0.01, 0.16]). This indicates that 6% of the variance in points is explained through the indirect effect of winning the Stanley Cup through games played.

Supplementary analyses 1

A main finding from the focal study reveals that winners get more opportunities than non-winners across their careers. Supplementary analyses were conducted to investigate possible mechanisms underlying opportunity differences and to extend our findings by testing another ancillary tenet of the Matthew Effect, which suggests that additional opportunities lead to resource gains. To investigate how status from winning an award may operate to create more resources for winners, we examined the timing of when winners acquire more resources and a possible mechanism of this acquisition. Given the expected mechanism of cumulative advantage operating over time, we wanted to examine when these opportunity differences may begin to arise and if they may be the result of formal work arrangements. Specifically, we wanted to determine (a) if winners procure enhanced opportunities (i.e., length of contract, total dollar amount of contract) immediately following their win and (b) if these advantages are formally incorporated into contracts. To examine these questions, we examined if winners were able to secure better contracts than finalists immediately following their win and loss, respectively. If winners receive better contracts immediately following their win, this would provide evidence for how cumulative advantage may operate over ones career. Namely, immediate advantages in resources codified in improved contractual agreements, which unfold over one’s career, and result in large differences at the conclusion of one’s career.

Participants and procedure

Consistent with our focal study, NHL players were used as participants and the Stanley Cup was considered a marker of affiliation-based external status. However, a new dataset had to be created. NHL contracts can be different lengths, varying from a single year to as many as 12 [30]. The Stanley Cup typically concludes in late May or early June. There is no guarantee that a player’s contract will be completed immediately after winning the Stanley Cup. For example, a player could sign a 7-year contract in 2020 and win the Stanley Cup in 2021. As such, the player was already committed to 6 more years with their team and winning the Stanley Cup would therefore not have any bearing on their contract. To examine if winners immediately benefited from winning the Stanley Cup, analyses must be completed on players whose contracts expired immediately after their Stanley Cup appearance. Very few players from the matched dataset fit this requirement, so we could not use the dataset utilized in the focal study.

As such, a new dataset was created for our first supplementary analysis. We chose to compare winners to finalists as finalists are a group that, beyond matched data, most closely resemble winners. We utilized an online hockey salary database [31] for player transaction information. We recorded all winning teams’ and finalist teams’ signings the year following their Stanley Cup win (i.e., winners) or Stanley Cup appearance (i.e., finalists). Specifically, we recorded individual player’s previous contracts and the newly negotiated contract’s length in years, the contract’s Average Annual Value (AAV; total compensation earned per year), and Total Salary (total career compensation). Of note, NHL salary data is not readily available prior to 2007. As such, data from Stanley Cup Finals from 2007–2019 was recorded. This resulted in contract data from 176 players (88 finalists, 88 winners).

Measures

Supplementary analyses 2

Supplementary Analyses 1 did not find any opportunity differences in the form of improved contracts immediately after affiliation-based external status acquisition. To follow up on the findings of the focal study, we extended our investigation to examine how affiliation-based external status may operate near the end of individuals’ careers. Professional sports leagues are highly competitive and there is a high rate of turnover from these leagues. As a consequence, a sign of player success is their career longevity. According to one analysis, in the NHL, bottom players play an average of 2 years, average players play 4.5 years, and the top players play an average of 12 years [32]. As a result, the ability to continue to play in the NHL represents a vital and foundational source of opportunity. In Supplementary Analyses 2, we sought to investigate if affiliation-based external status affords those with status more access to career extension opportunities.

In team settings, team output is observable and tangible, but it can be difficult to identify the distinct contributions of individual team members [33]. In team situations, the Matthew Effect suggests that high status team members tend get more of the credit for team performance, compared to less eminent team members [4]. More recent work has found that, in addition to getting more credit for good outcomes, more eminent members are punished less for bad outcomes when compared to less esteemed team members [33]. In light of the enhanced visibility of high-status team members, we investigated if status results in a player’s value at the end of their careers being differentially appraised. We inquire if status acts as a protective factor, such that the performance of those with status may be evaluated more favourably than those without status. As such, players with status may continue to get opportunities at the end of their careers in situations where players without status may not, as a result of more favourable evaluations.

Given the competitive nature of the NHL and limited roster space, teams are incentivized to remove unproductive players. Consequently, a marker of a player’s value is how much opportunity (i.e., ice time) a player receives by their teams within games (i.e., ATOI). Organizations allocate opportunities in a way that they feel gives them the best chance to be successful. As such, more valuable players typically get more ice time, or more time to influence the outcome of the game, than players that are not as valuable. Thus, as players’ playing time decreases in comparison to their average career ice time, we expect that there will be an increase in the likelihood of them exiting the league as they are being seen as less valuable by their organizations. However, if status acts as a protective factor and results in more favourable subjective evaluations of performance, we expect that players with status will not be subject to the same fate as those without status and decreased playing time will not result in an increased likelihood of exiting the league. Specifically, we examined if being a Stanley Cup winner alters the probability of exiting the league and ending ones’ career as a function of how much in game playing time they had the previous season.

Measures

Average Time on Ice (ATOI)

The average amount of time the player spent on the ice in the games he played (calculated by dividing each player’s seasonal time on ice by their total games played).