Persuaded Under Pressure: Evidence From the National Football League

Abstract

We exploit a natural experiment within each National Football League game, finding the first evidence in professional sports that referees succumb to the pressures of satisfying team personnel in the vicinity of possible violations. Using generalized additive models for binomial outcomes, we show that these sideline‐based differences in penalty rates, which are observed on common but influential penalties including pass interference and holding, peak near the centralized location of players and coaches on the sideline. With sizable interests in referee decisions, coaches and players often try to manipulate referee behavior with verbal and nonverbal communications; such actions appear to be persuasive. (JEL ZO, H3)

Modeling Player and Team Performance in Basketball

“…Tools from causal inference should also help us reason more soundly about questions of extrapolation, identifiability, uncertainty, and confounding, which are all ubiquitous in basketball. Based on our literature review, this need for causal thinking in sports remains largely unmet: there were few works which explicitly focused on causal and/or game theoretic analyses, with the exception of a handful in basketball (Skinner & Goldman, 2015, Sandholtz & Bornn, 2018 and in sports more broadly (Lopez, 2016, Yam & Lopez, 2019, Gauriot & Page, 2018.…”

Makeup calls in organizations: An application of justice to the study of bad calls.

“…In addition to moderating the direct effect of bad calls on the likelihood of makeup calls, we also expect outcome gravity and number of people impacted to have a first-stage moderation effect on the mediating role of guilt. Drawing from Jones (1991), situations characterized by greater outcome gravity and number of people impacted would be considered vivid and salient; that is, they should lead to stronger affective responses (see Griffin et al, 2016;Hall & Fincham, 2008;Mencl & May, 2009, 2016. These stronger affective responses subsequently impact cognitive and behavioral outcomes (e.g., moral intentions and decisions).…”

Policy Changes in Major League Baseball: Improved Agent Behavior and Ancillary Productivity Outcomes

“…For example, they are better at identifying the asymmetries of rounded strike zone edges of the zone, and avoid overfitting through cross-validation procedures to find the appropriate smoothing parameters. This estimation procedure has been used in past work investigating both MLB umpire and National Football League referee bias (Lopez 2016;Mills 2014;Tainsky, Mills, and Winfree 2015a). A full discussion of GAMs can be found in Gu and Wahba (1993) and Wood (2000Wood ( , 2003Wood ( , 2004Wood ( , 2006Wood ( , 2011, but I reprise the basic method developed in Mills (2014) for pitch location data here.…”