Career expectations and optimistic updating biases in minor league baseball players

https://doi.org/10.1016/j.jvb.2021.103615Get rights and content

Highlights

  • Athletes received career predictions from a C5.0 machine learning algorithm.

  • Most athletes updated their expectations when shown career information.

  • Asymmetric and contrary updating were observed.

  • Necessary optimism, reference group neglect, and affective information explain errors.

Abstract

Data on the likelihood of becoming a professional athlete are abundant and readily available, yet athletes consistently overestimate their chances of achieving the top levels of career success. Research is needed to examine whether athletes and others update their career expectations when seeing new information. In this study, minor league baseball players created a career tree estimating their probabilities of moving through the minor league system and then read personalized trees built by a C5.0 machine learning algorithm. After seeing the C5.0 trees, many players updated their expectations consistent with updating theory, especially when reevaluating their chances of being out of the system; however, there was evidence of asymmetric updating. Some acted opposite to what Bayesian reasoning would suggest. Analysis of the interview data reveals three themes that explain asymmetric and contrary updating. Players believed optimism is necessary for their baseball career, they neglected their reference group, and they saw information as possessing affective qualities. Using these three themes caused athletes to ignore some information and, occasionally, circumvent the updating process altogether.

Introduction

Researchers have found that athletes have unrealistic expectations of becoming professionals (Kennedy & Dimick, 1987; NCAA, 2019; Sailes, 1998). Athletes with unrealistic career expectations enter development systems with low pay (Pifer et al., 2020), make errors when negotiating and signing professional contracts (McCann, 2006), and avoid career planning (Park et al., 2012; Peptitpas et al., 1990). Athletes who experience unprepared and unplanned retirements suffer from anxiety and depression, they feel anger, disappointment, and regret, lack autonomy in their career decisions, and are less satisfied with their lives (Knights et al., 2019; Perna et al., 1999). Athletes are not the only people with unrealistic career expectations. Students frequently overestimate their salaries after graduation (Jones et al., 2020; Shepperd et al., 1996), which leads students to pick different majors than if their expectations were accurate (Wiswall & Zafar, 2015). Therefore, it is imperative to help athletes and others develop accurate career expectations.

Information about careers, such as probabilities published by the National Collegiate Athletic Association on the chances of becoming a professional athlete, should help athletes develop realistic expectations (Wiswall & Zafar, 2015). However, there has been little research into how athletes perceive career information, and there is no theory explaining how athletes update their expectations when shown new information. If athletes acted consistent with Bayesian theory, the leading normative theory for updating beliefs, they would alter their expectations when faced with new information (Silver, 2012). However, researchers have demonstrated that people make consistent errors when processing new information (Eli & Rao, 2012; Garrett & Sharot, 2017; Sharot et al., 2011). Athletes’ updating practices might differ from normative recommendations, causing them to maintain unrealistic expectations even though they have the information needed to develop accurate expectations. Thus, the purpose of this study is to examine how athletes update their expectations when shown career information.

Although this study focuses on athletes, we expect it will advance research on careers more generally due to the importance of expectations and information in leading career theories. For example, outcome expectations are one of the central features of social cognitive career theories (Ireland & Lent, 2018; Lent & Brown, 2013), and expectations of success are one of the fundamental beliefs in the expectancy-value theory of achievement choices (Gao & Eccles, 2020; Wigfield & Eccles, 2000). Information about the world of work is also one of the critical ingredients of effective career choice interventions (Brown et al., 2003). However, Lent and Brown (2020) explained that career decision-making models have tended to assume people are rational decision-makers when there is growing evidence suggesting widespread cognitive biases and other errors in decision-making. Thus, it is likely that many people have unrealistic expectations about their careers, and it is vital to study whether they use information rationally when updating their expectations. Professional athletes are ideal for exploring these questions because they have clear, incremental career pathways that enable researchers to measure expectations. Researchers can also use public information on performance and career outcomes to determine whether athletes have accurate expectations (see also Richardson & McKenna, 2020).

To examine how athletes update their expectations when shown career information, we elicited career expectations from minor league baseball players. We then showed them personalized career predictions created by a C5.0 machine learning algorithm and data on every player drafted between 2003 and 2011. By analyzing athletes’ responses to the C5.0 data, this study makes the following contributions. First, we examine how athletes’ updating behaviors differ from recommendations of normative theory. Second, we use qualitative data to propose theoretical explanations for why athletes incompletely update. Third, we show that some athletes violate normative principles when updating, and we attempt to explain why. These theoretical contributions will serve as the groundwork for a better understanding of unrealistic career expectations, which will inform future research on careers and help counselors and other career development professionals intervene to encourage athletes and others to think more accurately about their careers.

If people were rational “econs,” they would update their career expectations using the practices supported by Bayesian statistical theory (Kahneman, 2011; Silver, 2012). Bayesian principles recommend estimating prior probabilities and then updating those probabilities when encountering new information. More specifically, Bayes’ theorem describes how to calculate posterior probabilities, from priors, given a new event. If used correctly, two people using Bayesian updating will converge on the same expectations about the future, irrespective of how their expectations differed, so long as they receive the same information (Silver, 2012). Bayesian theory is an ideal standard for updating practices because it describes how to use available information as effectively as possible to develop the most accurate expectations.

Applied to becoming a professional athlete, Bayesian athletes would estimate a prior probability of making it and use any events to estimate new posterior probabilities. For example, being selected by a travel team, being ranked by a prospect site, or seeing a personalized career prediction would provide information to update a prior probability to estimate a new posterior probability of becoming a professional. Thus, according to Bayesian reasoning (e.g., Kahneman, 2011), athletes should start with base rate expectations and incrementally adjust their expectations as they move further through the development pathway. Bayesian reasoning is indifferent to whether athletes start with high or low expectations, so long as they update in the direction implied by new information.

However, researchers have demonstrated that people do not always update their beliefs consistent with normative theory. Massey et al. (2011) found that fans’ predictions of National Football League games remained optimistically biased over a season of games, despite accuracy incentives and extensive feedback. Although fans’ predictions became more accurate over the season as they learned from experience (better calibration), they still exhibited optimism bias (consistent upward and downward errors). Massey et al. (2011) further showed that optimistic biases were explained mainly by fans’ desires to see their favorite teams win over others. Therefore, although people can learn and update their beliefs, they also consistently fall short of rational models, which is likely to happen in career updating.

One theory explaining why people make updating errors is that they are optimistically biased when updating, which leads them to update their beliefs to a greater extent when receiving information perceived as good than when receiving information perceived as bad (Eli & Rao, 2012; Garrett & Sharot, 2017; Sharot et al., 2011). This asymmetry can lead to the persistence of positive beliefs regarding one's future, known as unrealistic optimism (Weinstein, 1980), which has been demonstrated for many life outcomes, including outcomes that people could mitigate with behavioral changes, such as earning a score on an exam or dying from lung cancer or heart disease (Shepperd et al., 2013). Given the persistence of unrealistic optimism in human psychology, it is possible that athletes and other workers asymmetrically update their career expectations when shown new career information.

Athletes’ unrealistically optimistic expectations might also be linked to dispositional optimism and optimism-based mental skills that are favored in sport and other performance contexts. Optimism, and related ideas, such as confidence, hope, and self-belief, are woven into the fabric of sports culture because of widespread scientific and anecdotal evidence that they help performance (Curry et al., 1997; Grove & Heard, 1997). Sport psychologists have recommended teaching optimism skills to improve athletes’ mental toughness and performance (Nicholls et al., 2008). Elite and sub-elite athletes report higher dispositional optimism than non-athletes and less accomplished athletes (Bleichrodt et al., 2018; Venne et al., 2006).

The health and performance benefits of dispositional optimism are also well recognized outside of sport contexts. Optimists persevere, work longer hours, are healthier, and live longer (Puri & Robinson, 2007; Sharot, 2011; Taylor and Brown, 1988, Taylor and Brown, 1994). Indeed, given the widespread prevalence of optimism and other positive illusions among humans, researchers have suggested they might have evolutionary advantages (Johnson & Fowler, 2011). Although dispositional and skill-based optimism have benefits, an unintended consequence is that athletes and other performers might have unrealistically optimistic attitudes toward their careers and new career information. However, we are not aware of any empirical work that has examined how athletes update their expectations.

In summary, although Bayesian updating is the ideal framework for updating expectations about the future, humans have many cognitive and emotional biases that can cause them to err when updating, and these biases might be pronounced and observable among athletes. To examine these topics, we asked the following research questions:

  • RQ1: How do athletes update their expectations when shown career information?

  • RQ2: In what ways do athletes’ updating behaviors differ from practices suggested by normative theory?

  • RQ3: What reasons, justifications, or explanations do athletes give that might explain why their updating behaviors differ from normative theory?

Section snippets

Method

We conducted interviews with minor league baseball players using a career tree protocol to examine how athletes respond to career information. We chose minor league baseball players as participants in this study for three reasons. First, athletes advance through the minor league farm system one level at a time, making it ideal for measuring athletes’ career expectations. Second, since there are abundant data on player performance and career trajectories, we could show players career predictions

Results

Most players’ prior probabilities required updating according to our C5.0 predictions. Across three career tree measures, only 10% to 19% of players made predictions within 5% of our C5.0 model (Table 1). The remaining athletes made predictions more than 5% above or below the C5.0 data and, therefore, should be expected to use new information to update their expectations. Moreover, players consistently made optimistic predictions. Most players overestimated their chances of making it to the

Discussion

The purpose of this study was to examine how athletes update their expectations when shown career information. Our findings show that most athletes updated their career trees when shown personalized career predictions created by a C5.0 machine learning algorithm. Athletes’ updating behavior often followed general updating principles. For example, most athletes who underestimated their chances of being out of baseball increased their chances of being out after seeing their C5.0 tree. However,

CRediT authorship contribution statement

Christopher M. McLeod: Conceptualization, Methodology, Validation, Investigation, Writing – original draft, Writing – review & editing, Visualization, Project administration. N. David Pifer: Methodology, Software, Validation, Formal analysis, Writing – review & editing. Emily P. Plunkett: Validation, Writing – review & editing.

Declaration of competing interest

Chris is an unpaid advisor to More Than Baseball, an organization created to help minor league baseball players.

Acknowledgements

The authors thank Michael Sagas and two reviewers for their comments and feedback.

References (47)

  • R. Berengui et al.

    Optimism and burnout in competitive sport

    Scientific Research: Psychology

    (2013)
  • H. Bleichrodt et al.

    The risk attitudes of professional athletes: optimism and success are related

    Decision

    (2018)
  • C. Camerer et al.

    Overconfidence and excess entry: an experimental approach

    The American Economic Review

    (1999)
  • L.A. Curry et al.

    Role of hope in academic and sport achievement

    Journal of Personality and Social Psychology

    (1997)
  • D. Eli et al.

    The good news-bad news effect: asymmetric processing of objective information about yourself

    American Economic Journal: Microeconomics

    (2012)
  • L.G. Epstein et al.

    Non-Bayesian learning

    The BE Journal of Theoretical Economics

    (2010)
  • J.R. Grove et al.

    Optimism and sport confidence as correlates of slump-related coping among athletes

    The Sport Psychologist

    (1997)
  • A.F. Hayes et al.

    Answering the call for a standard reliability measure for coding data

    Communication Methods and Measures

    (2007)
  • D.D.P. Johnson et al.

    The evolution of overconfidence

    Nature

    (2011)
  • S. Jones et al.

    Misinformed, mismatched, or misled? Explaining the gap between expected and realized graduate earnings in Mozambique

  • D. Kahneman

    Thinking, fast and slow

    (2011)
  • S.R. Kennedy et al.

    Career maturity and professional sports expectations of college football and basketball players

    Journal of College Student Personnel

    (1987)
  • S. Knights et al.

    The end of a professional sport career: Ensuring a positive transition

    Journal of Sport Management

    (2019)
  • Cited by (0)

    View full text