Navigating Darkness: College Athlete Suicide, Support Systems, and Shadows of Depression

Authors: Matt Moore, Ph. D, MSW 1, Anne M. W. Kelly, Ph. D 2, Lana Loken, Ed. D. ATC 2, Mastano N. Dzimbiri, MS 1, Payton Bennett, student

Corresponding Author:

Matt Moore, Ph. D, MSW
Chair and Faculty, Family Science and Social Work Department
Miami University
501 E. High Street

Coaches’ Perspectives of the Influence of Safe Sport-Related Education 


Purpose: An increase in mental health concerns and suicide among young adults led to a sharpened research focus on suicide and college athletes. In this study, we investigated the relationship between college athletes’ risk of depression, suicidality, and their support system and whether preventing suicide deaths requires identification of commonly cited risk factors. Methods: Voluntary college athletes aged 18-years-old or older and attending an NAIA member institution participated in the study (n = 361). They completed a web-based instrument that consisted of the following: (1) demographic questionnaire, (2) Patient Health Questionnaire (PHQ-9), (3) Berlin Social Support Scale, and (4) Columbia Suicide Severity Rating Scale. Results: Between 5-18% of college athletes responded affirmatively to one of the questions asking about suicidality. There was a significant moderate negative correlation between the suicide predictor and the PHQ-9 score and significant weak positive correlations between the suicide predictor and perceived emotional support and between the suicide predictor and perceived instrumental support. Conclusion: This study identified findings that might be useful to practitioners and opened new lines for future research. Applications in Sport: College athletic programs and university counseling centers are poised to enhance our understanding of student-athletes’ suicidal distress and how to respond by making use of qualitative research methods. We strongly recommend adopting this strategy to address depression and suicidal ideation.

Keywords: prevention, student-athletes, mental health, risk factors

Despite growing openness about mental health struggles, a disparity still exists between physical and mental health (Gorczynski et al., 2023; Moore et al., 2022), fostering stigma and hindering help-seeking behavior (Moore, 2017), particularly among college students (Centers for Disease Control and Prevention [CDC], 2021). While mental health diagnoses in the college student population is a longstanding challenge, the COVID-19 pandemic increased stressors placed on the college student population leading to increased risks (Gupta & Agrawal, 2021; MacDonald & Neville, 2023).

According to the CDC (2021), mental health concerns and suicidal thoughts are increasing for youth and young adults. Forty percent of those surveyed showed signs and symptoms of depression and 20% said they had thoughts of suicide. These trends are similar to studies on college student mental health and suicidality (Barclay et al., 2023; Schmiedehaus et al., 2023). According to the Substance Abuse and Mental Health Services Administration (SAMHSA, 2017) individuals aged 18-25 reported a 3% increase in major depressive episodes from 2015-2017. Additionally,18.9% of individuals 18 and above reported experiencing a mental illness in the past year, with 7.5% reporting a serious mental health illness (SAMHSA, 2017). A second SAMHSA (2021) study found 33.7% of individuals aged 18-25 reported a mental illness and 11.4% reported a serious mental illness.
In addition to concerns about serious mental health illness, SAMHSA (2021) found an increase in rates of suicidal behavior. Specifically, 10.5% reported having serious thoughts of suicide, 3.7% created a suicide plan, and 1.9% attempted suicide. Research by Rosenthal et al. (2023) found higher rates with 13.7% of college students reporting suicide ideation, 7.6% making a suicide plan, and 3.2% reporting at least one suicide attempt. In 2021 suicide became the leading cause of death for those aged 20-24 (CDC, 2023).
One subset of the college student population is college athletes. Recently, discussion of their mental health increased. Researchers attempted to explore the intersectional identity of student athletes and the effect that this role strain may have on mental health (Gorczynski et al., 2023; Moore et al., 2022). Quantifying mental health and suicide risk in this group is challenging, with conflicting results on the link between depression, support systems, and suicide. Many researchers see sport participation as a protective factor for mental health risk due to the social support provided by the team (Hui et al., 2023; Sullivan et al., 2020). But additional pressures like failure to successfully compete or live up to expectations, loss of social structure due to injury or retirement from sport, or time demands of the sport in addition to being a college student can increase the risk (Moore, 2017; Moore et al., 2022). This study builds upon existing research by looking more closely at the relationship between a college athletes’ risk of depression, suicidality, and their support system.

College Athletes and Depression
According to the American Psychological Association (2020), depression is one of the most common mental health disorders in the United States. Depression might include emotional, cognitive, physical, and/or behavioral symptoms and is best understood on a continuum of severity, rather than either present or not present. Findings amongst college athletes demonstrate that depression rates align with rates of the general population of college students (hovering around 25%) (Prinz et al., 2016; Wolanin et al., 2016), and some revealed that athletes have higher rates of depression (over 30%) than the general population (Cox, 2015). While many studies find similar rates between college athletes and their non-athlete peers, others show participation in college athletics can decrease one’s risk for depression (Banu, 2019; Salehioan et al., 2012).
Although some research shows athletic participation may protect against mental illness, there is still reason for concern for college athletes. A current study by the National Collegiate Athletic Association (NCAA, 2022) surveyed almost 10,000 NCAA athletes from all three competitive division levels. Results showed athletes of all competition levels demonstrated elevated levels of mental exhaustion, anxiety, and depression. These levels were nearly two times higher than pre-pandemic levels. The top three factors negatively affecting mental health were academic worries (44%), planning for the future (37%), and financial worries (26%). Only 50% of college athletes believed mental health was a priority for their athletic department, 33% of college athletes did not know where to go to seek mental health services, and as many as 17% of college athletes reported feeling hopeless.

College Athletes and Suicide
Suicide risk in athletes is difficult to determine due to underreporting and misclassification of many sudden deaths. Over the past two decades the NCAA attempted to determine the risk of suicide specific to college athletes. Rao et al. (2015) reported that 7.3% of all athlete deaths were suicides, making suicide the fourth leading cause of death for college athletes. Previously, Miller and Hoffman (2009) found approximately 5% of student-athletes contemplated suicide. Much like research on college athlete depression, some research demonstrates sport protects against suicidality (Maron et al., 2014). This study’s findings highlight the importance of promoting participation in diverse sporting activities among college students given that engaging in such activities safeguards against depression and suicidal ideation by nurturing self-esteem and bolstering social support.

College Athletes and Social Support
The discrepancy in the literature may be accounted for by the supports that are available to college athletes and their willingness to seek such supports (Sullivan et al., 2020). One of the most discussed supports is the team environment. Sullivan et al. (2020) analyzed the effects of social supports on depressive symptoms in college athletes. They found emotional support from teammates, family, and friends was correlated with a decrease in depressive symptoms. Other more formal or instrumental supports that reduced depression included the availability of tutoring and health services, including mental health providers with specialization with athletes.
Social support has not been as extensively studied in the college athlete population. Studies show links between social support and burnout as well as social support and overall wellbeing in college athletes (Defreese & Smith, 2014). Research identified social support as an important component in allowing athletes to balance school and athletics (Carter-Francique, 2015). Many college athletes have strong social support networks naturally, such as relationships with teammates, coaches, medical staff, and other resources provided by the athletic department (Armstrong & Oomen-Early, 2009). They also have supportive relationships, such as family and friends, outside of athletics.
Despite knowledge of these available supports and benefits they offer college athletes, exploring the utilization of built-in athletic supports and personal supports unique to an individual athlete remains understudied. Much of the research tends to oversimplify social support. Due to its dynamic and complex nature, social support among college athletes merits further investigation. Research has not examined the differences in the type of perceived social support in collegiate athletics as it relates to levels of depressive symptoms and suicidality.

Present Study
Overall, the research on mental health issues, including depression and suicide in collegiate athletes is inconclusive. More research is needed to determine what factors put athletes at risk for severe mental health concerns and suicide. The purpose of this study was to investigate whether there is a relationship between levels of depression and suicide risk and levels of social support among National Association of Intercollegiate Athletics (NAIA) college athletes. The NAIA does not have data available on connectedness between depression, social support, and suicide.



Research Design
The current exploratory study utilized a cross-sectional, web-based survey design to gather data from NAIA college athletes. Considering the size of the NAIA student-athlete population, confidence level, confidence intervals, statistical test, and statistical power, the minimum sample for this study was 47 college athletes (Faul et al., 2007). Researchers identified athletic trainers through the NAIA database to establish contact information. Athletic trainers provided survey information to their assigned college athletes. This approach was successful in other NAIA research efforts (Moore & Abbe, 2021).

The exploratory study utilized a stratified random sampling procedure to identify college athlete participants. Researchers divided the NAIA college athlete population into subgroups, or strata, based on sports available throughout the NAIA. This included a stratum for each of the 17 sports with separate stratum for each gender that participates in a sport. Next, researchers identified NAIA member institutions that participated in each of the 17 sports. Each institution participating in a sport received a random number. Researchers selected random numbers to identify the member institutions that would participate in the survey from each sport. This approach ensured all member institutions participating in various sports had an equal opportunity for inclusion.

Voluntary college athletes aged 18-years-old or older and attending an NAIA member institution participated in the study (n = 361). Most participants were 18-21 years old (53.5%, 46.5% indicated being over the age of 21). Survey participants were primarily juniors (30.7%, 23.8% sophomores, 23.1% first years, 22.1% seniors of graduate students). More women completed the survey (59.8%, 40.2% men). Most participants who reported race/ethnicity were White/Caucasian (55.4%, 21.9% Hispanic or Latino, 14.9% Black or African American, 6.6% multiracial, 1.2% from other groups).

Table 1.

NAIA Institutional Demographic Information

University Demographic%
Faith Based62.9%
Non-Faith Based37.1%

Participants recorded which NAIA athletic team they were primarily affiliated with (20.2% baseball, 19.9% soccer, 12.5% track volleyball, 8.0% softball, 6.4% cross country, 6.1% basketball, with all other sports being under 5% each [e.g., football, bowling, cheer, dance, track and field, swimming and diving, golf, tennis, and lacrosse]). Participants were further examined regarding NAIA college/university demographics (See Table 1). Participants also responded to whether or not they receiving mental health training from their college of university before participating in sport. The largest majority (n = 229, 63.7%) indicated they did not receive such training. The other 36.3% (n= 132) indicated they did receive some form of training.
[Insert Table One]

Measures and Instruments

College athletes completed a web-based instrument that consisted of the following: (1) demographic questionnaire (see above demographics), (2) Patient Health Questionnaire (PHQ-9; Kroenke et al., 1999), (3) Berlin Social Support Scale (BSSS; Shulz & Schwarzer, 2003), and (4) the Columbia Suicide Severity Rating Scale (C-SSRS; Posner et al., 2011). 

Patient Health Questionnaire (PHQ-9)
The PHQ-9 is a self-administered version of the PRIME-MD diagnostic instrument for common mental disorders (Kroenke et al., 2001). It is used to make criteria-based diagnoses of depressive and other mental disorders commonly encountered in primary care. This is a 9-item depression module upon which the diagnosis of Diagnostic and Statistical Manual (DSM) depressive disorders is based. Reliability and validity of the tool have indicated it has sound psychometric properties. Internal consistency of the PHQ-9 has been shown to be high (American Psychological Association, 2020). There is precedent for using the PHQ-9 in research with college athletes (DaCosta et al., 2020; LoGalbo et al., 2022).

Berlin Social Support Scale (BSSS)
The researchers measured the degree of emotional and tangible support using the BSSS (Schulz & Schwarzer, 2003). This scale measured perceived emotional and instrumental supports, need for support, and support seeking. There are 17 items on the BSSS that are answered using a five-point Likert scale with endpoints “1 = Strongly Disagree” and “4 = Strongly Agree.” The researchers used a mean score for each of the subscales (perceived emotional support, perceived instrumental support, need for support, and support seeking). The scale has a Cronbach’s alpha of 0.83 for perceived social support, 0.63 for need for support, and 0.83 for support seeking (DiMillo et al., 2017). The scale has a prior history of use within college athletics (Sullivan et al., 2020)

Columbia Suicide Severity Rating Scale (C-SSRS)
The C-SSRS was developed by researchers from Columbia, Pennsylvania, and Pittsburgh Universities to evaluate suicidal ideation and behavior (Posner et al., 2011). The scale provides a brief assessment of severity and intensity of suicidal ideation, suicidal behavior, and lethality (Syndergaard et al., 2023). The screener version used in this study consisted of six “yes” or “no” questions. Based on participant responses to the six questions, participants were considered low, moderate, or high risk. The C-SSRS has excellent internal consistency (α = 0.95). Principal components analysis revealed a two-factor solution, accounting for 65.3% of the variance across items (Madan et al., 2016). There is limited research on the use of the C-SSRS with the athlete population (Costanza et al., 2021).

Data Collection
Researchers contacted the athletic training staff at all sampled NAIA member institutions. Athletic training staff received the list of teams from their institution for inclusion in data collection. Researchers provided athletic training staff detailed instructions for data collection and a copy of the informed consent. Athletic training staff distributed the electronic survey to their college athletes. College athletes were able to opt-out of the survey at any time. The survey took approximately 15-20 minutes to complete. Researchers recorded survey results into a statistical software program (SPSS 28) on a secure, private platform.

Data Analysis
Researchers utilized descriptive statistics to provide details about the sample and overall survey results. Researchers used inferential statistics to infer information from the sample data to the overall NAIA student-athlete population.

To investigate the first research objective, an initial correlation analysis was conducted to examine whether having any safe sport training was related to increases in coaching outcomes. The safe sport training variable was transformed so that coaches who answered “yes” to completing any of the safe sport training courses were coded as 1 and coaches who had answered “no” to completing all the safe sport training courses were coded as 0 (i.e., no SS training=0, any SS training=1). This variable was included in a correlation analysis with all coaching outcomes: knowledge & confidence, safe sport stress, stress over athlete well-being, and efficacy to support others. To investigate the second research objective, four separate linear regression models were constructed with the sum of completed safe sport training courses (range =1-12) as the independent variable, and the following coaching outcomes as respective dependent variables: knowledge & confidence, safe sport stress, stress about athlete well-being, and efficacy to support others. In all four models, the coaching context, whether training was required (0=no, 1=yes), and whether training was free (0=no, 1=yes) were included as covariates. To address the third research objective, ANOVAs were conducted with individual safe sport courses as independent variables, and the following coaching outcomes as dependent variables: knowledge & confidence, efficacy to support others, safe sport stress, stress about athlete well-being and efficacy to support others. All analyses were conducted using IBM SPSS Statistics (Version 28) (20).


Descriptive Statistics
College athletes answered each item from the C-SSRS. Descriptive findings from this scale indicated that 18.3% of participants wished to be dead, 18,3% had non-specific active suicidal thoughts, 13.6% had active suicidal ideation without intent to act, 6.1% had active suicidal ideation with some intent to act, and 5.0% had active suicidal ideation with a specific plan and intent to act. Of the 361 college athlete respondents, 25.8% answers “yes” to at least one of the questions on the scale.

College athletes completed the PHQ-9 as a brief screening tool for potential depressive symptoms. Results of the PHQ-9 and the percent of athletes at risk of depression for each item can be found in Table 2.

Table 2. PHQ-9 Scores for NAIA College Athletes

QuestionMean (SD) (% At Risk)
Little interest or pleasure in doing things?1.81 (0.91) (22.1%)
Feeling down, depressed, or hopeless?1.68 (0.81) (14.1%)
Trouble falling asleep or sleeping too much?2.06 (1.05) (30.2%)
Feeling tired or having little energy?2.17 (0.92) (29.1%)
Poor appetite or overeating?1.81 (0.96) (21.3%)
Feeling bad about yourself?1.75 (0.93) (18.6%)
Trouble concentrating on things?1.69 (0.96) (17.2%)
Moving or speaking so slowly that people could have notice? Or more fidgety and restless than usual?1.34 (0.69) (7.8%)
Thoughts that you would be better off dead?1.21 (0.53) (4.1%)

Evaluation of Assumptions

College athletes also completed the BSSS. Results of the BSSS and the percent of athletes at risk of limited social support in various areas can be found in Table 3. These are only the scale items where there were significant concerns about perceived emotional support, perceived instrumental support, need for support, and support seeking.

BSSS Scores for NAIA College Athletes

QuestionMean (SD) (% At Risk)
Whenever I am not feeling well, other people show me that they are fond of me? 3.14 (0.82) (17.2%)
When everything becomes too much for me to handle, others are there to help me?3.21 (0.83) (18.3%)
I get along best without any outside help?2.48 (0.81) (48.7%)
In critical situations, I prefer to ask others for their advice?3.00 (0.79) (23.0%)
Whenever I am down, I look for someone to cheer me up again?2.51 (0.89) (49.6%)
When I am worried, I reach out to someone to talk to?2.69 (0.93) (38.2%)
Whenever I need help, I ask for it.2.70 (0.96) (39%)

Researchers used correlation analysis to assess the relationship between a college student-athletes predictor of suicide with their score on the PHQ-9, perceived emotional support, perceived instrumental support, level of needed support, level of support sought, and mental health training.

Prior to conducting the analysis, researchers generated several statistics and graphs to examine the tests of assumption, including level of measurement, related pairs, absence of outliers, and linearity.

Results of the Correlational Analysis
Researchers computed a Pearson product-moment correlation coefficient to assess the relationship between a college student-athletes suicide predictor and their PHQ-9 score, perceived emotional support, perceived instrumental support, level of needed support, and level of support sought. There was a significant (p < 0.001) moderate negative correlation, r = -.462, N = 361 between the suicide predictor and score on the PHQ-9. There was a significant (p < 0.001) weak positive correlation, r = .236, N = 361 between the suicide predictor and perceived emotional support. A similar significant (p < 0.001) weak positive correlation, r = .255, N = 361 between suicide predictor and perceived instrumental support. A college student-athlete’s exposure to mental health training, perceived level of needed support, and level of support sought did not appear to be suicide predictors.


In this study, we investigated whether preventing suicide deaths requires the identification of factors that are associated with people’s risk of suicidal behavior. Commonly cited risk factors for suicidal thoughts and behaviors are depression and inadequate support. Association between major depressive disorder (MDD) and suicide attempts or ideation has been well-documented. Accordingly, depression has been considered a necessary or sufficient cause of suicidal thoughts. But much is unknown about the characteristics that increase suicide risk among people living with depression (Bradvik, 2018). Many mechanisms could play a role in suicidal behavior among people with MDD, and, although suicidal behavior occurs among people with major depressive disorder, depression is not necessarily a useful tool for understanding the complexity of suicide (Orsolini et al., 2020).

Most people with depression do not attempt suicide. Diagnosis of MDD requires a simultaneous presentation of several specific symptoms. Approximately, 17 million American adults will have symptoms of MDD each year, but only around 45,000-50,000 Americans will die by suicide during that same time. Considered independently of other risk factors, MDD may put one at greater risk, meaning that those with this disorder are more likely than those without it to die by suicide. But still very few of those with MDD will go on to die by suicide; reliance on depression to predict suicidality is inadvisable. This is supported by Ribeiro et al. (2018), who reviewed existing literature on the subject and showed that although depressive symptoms were reported to confer risk of suicidality, the effects were weaker than expected.

Melhem et al. (2019) demonstrated that the most severe depressive symptoms and variability over time were the only predictors of suicide attempt in young adults, especially when combined with other factors (e.g., childhood abuse, history of attempt, substance use disorder, and parental attempt). But prediction was marginally better than chance, perhaps because suicidal risk varies during a psychiatric illness and may be linked to other factors that appear during depressive episodes. Orsolini et al. (2020) showed that anxiety disorders co-occurring with MDD are among the main predictors of attempts. Several factors interact and contribute to suicidal behavior and death by suicide. These may include major depressive disorder, but interactions with other factors, such as genetic vulnerability, stress, psychiatric comorbidities, and social aspects need to be evaluated to improve prevention (Orsolini et al., 2020).
Results from our research showed a moderate negative correlation between the suicide predictor and score on the PHQ-9, challenging the assumption that depression is a necessary or sufficient cause of suicidal thoughts. This lends support to the idea that traditional risk factors can be problematic and that their predictive value has not improved over the past 50 years (Franklin et al., 2017; Fortune & Hetrick, 2022).

Bradvik (2018) also acknowledged that depression is related to suicidal ideation and attempt but is not a good predictor. Bradvik (2018) pointed to results from the Australian Rural Mental Health Study in which only 364 out of 1051 respondents reported life-time depression. Of those 364 respondents, 48% reported life-time suicidal ideation and 16% reported a suicide attempt. Gender, age of depression onset, and possibly psychiatric comorbidities were somewhat predictive of suicide behavior, but no other predictive factors were revealed. These results were echoed by Melhem et al. (2019).

The limits of risk factors to accurately predict suicide is further strengthened by our finding that an increase in emotional social support was weakly associated with an increase in suicide risk, contradicting earlier research that showed suicidal distress was worse when emotional social support was low (Ayub, 2015; Otsuki et al., 2019). Similarly, instrumental social support (i.e., support that helps people with practical tasks) was weakly associated with suicide risk, contradicting findings from Otsuki et al. (2019).
After a concussion, athletes experience a range of psychological symptoms, with depression and anxiety being among the most reported (Kontos et al., 2012). Symptoms can include loss of interest in activities that were once enjoyable, persistent sadness, physical and mental fatigue, and changes in sleep patterns. These negative outcomes may be more pronounced in athletes who attach a great degree of importance to the athlete’s role in relation to other activities (Brewer et al., 1993; Raedeke & Smith, 2001) and can be made worse by changes in lifestyle, the loss of social support that team members provided, and even personality traits. One such trait is maladaptive perfectionism.
Maladaptive perfectionists are overly critical of mistakes. They strive for excessively high and ultimately unobtainable goals. This usually results in failure, which can be painful, especially for athletes with maladaptive perfectionism, who may lack resilience to bounce back from stressful experiences. This unhealthy perfectionism is associated with higher levels of depressive symptoms (Egan et al., 2011; Olmedilla et al., 2022). Additionally, perfectionists can struggle with time management, not setting realistic timelines for getting things done or because they are paralyzed by the prospect of failure. Time management is one of the most difficult aspects of participating in college sports (Rothschild-Checroune et al., 2013).

Taken together, injury and concussion, personality traits (e.g., maladaptive perfectionism), and external factors (e.g., time constraints) can contribute to negative mental health outcomes among student-athletes and may increase suicidal distress. College athletic programs and university counseling centers are poised to improve our understanding of the nature of suicidal distress among student-athletes face and how to respond by making use of qualitative research methods, which we recommend. We urge university administrators to dedicate more resources to building and integrating academic and co-curricular resilience programs into their campuses and rely less on risk assessment that focuses on commonly cited factors (e.g., depression) to predict suicide.

Study Limitations
While efforts were made to decrease discomfort with the survey, it is possible college athletes felt pressure to respond in particular ways out of personal and/or athletic concerns. This study also relied upon self-reported data. Without having the ability to verify participant responses, there was no way of knowing the legitimacy or honesty of participants’ responses. The study was unable to control the multiple covariates or confounding variables that influence a college suicidality and mental health. Finally, our study lacked a detailed exploration of how specific socio-demographic characteristics, such as race, gender, and class status, might influence suicidal ideation and other risk behaviors among college athletes.

Future Research
The complex interplay between core risk factors in individuals and heightened suicide risk among athletes necessitates further exploration. Future research should focus on understanding the repercussions of escalated demands on athletes’ mental well-being, particularly the impact of significant situational factors such as career-ending injuries on their mental health and suicide vulnerability. Additionally, there is a need to delve into the connection between suicide rates, race, and gender among collegiate students for a more comprehensive understanding of these dynamics.

This study examined the relationship between college athletes’ risk of depression, suicidality, and their support system and whether preventing suicide deaths requires identification of commonly cited risk factor. The results are quite different from previous research findings, revealing a moderate negative correlation between the suicide predictor and scores on the PHQ-9, adding nuance to the presumption that depression is either a necessary or sufficient factor for the emergence of suicidal thoughts. College athletic programs and university counseling centers are poised to enhance our understanding of student-athletes’ suicidal distress and how to respond by making use of qualitative research methods. We strongly recommend adopting this strategy to address depression and suicidal ideation.

Applications in Sport
Studying suicide in college sports has practical applications that can help improve the well-being and safety of college athletes. By examining the factors that contribute to suicidal ideation and behavior in college sports, researchers and practitioners can develop targeted interventions and support systems to address mental health challenges. For instance, such studies may lead to the creation of tailored mental health resources for student-athletes, including counseling services and peer support networks. Furthermore, understanding the unique stressors faced by student-athletes, such as performance pressure and balancing academics with athletics, can inform the design of preventative measures such as stress management and resilience training programs. Additionally, awareness campaigns can be created to destigmatize mental health struggles in sports, encouraging athletes to seek help when needed. Overall, studying suicide in college sports can lead to a safer and more supportive environment for student-athletes, promoting their overall health and success.


1. American Psychological Association. (2020). Patient Health Questionnaire.,been%20shown%20to%20be%2high.
2. Armstrong, S., & Oomen-Early, J. (2009). Social connectedness, self-esteem, and depression symptomatology among collegiate athletes versus non-athletes. Journal of American College Health, 57(5), 521-526.
3. Ayub, N. (2015). Predicting suicide ideation through intrapersonal and interpersonal factors: The interplay of Big-Five personality traits and social support. Personality and Mental Health, 9(4), 308-318.
4. Banu, P. R. (2019). Depression among athletes and non-athletes university students. International Journal of Social Sciences, 7(5-III), 1565-1569.
5. Barclay, N., Kelley, K. A., Brausch, A. M., Muehlenkam, J. J., & Nadorff, M. R. (2023). Changes in the suicide risk behaviors of American college students over time: An analysis of three universities. Suicide & Life-Threatening Behavior, 53(5), 764-775.
6. Bradvik, L. (2018). Suicide risk and mental disorders. International Journal of Environmental Research and Public Health, 15(9), 2028.
7. Brewer, B. W., Van Raalte, J. L., & Linder, D. E. (1993). Athletic identity: Hercules’ muscles or Achilles heel? International Journal of Sport Psychology, 24(2), 237–254.
8. Carter-Francique, A. (2015). Examining the value of social capital and social support for Black student-athletes’ academic success. Journal of African American Studies, 19(2), 157-177.
9. Centers for Disease Control and Prevention. (2023). Youth risk behavior survey: Data summary and trends report.
10. Centers for Disease Control and Prevention (2021). Suicide data and statistics.
11. Costanza, A., Radomska, M., Zenga, F., Amerio, A., Aguglia, A., Serafini, G., Amore, M., Beradelli, I., Ojio, Y., & Nguyen, K. (2021). Severe suicidality in athletes with chronic traumatic encephalopathy: A case series and overview on putative etiopathogenetic mechanisms. International Journal of Environmental Research and Public Health, 18(3), 876.
12. DaCosta, A., Grueninger, K., Hurt, S., Crane, A., Webbe, F. M., & LoGalbo, A. (2020). A-11 comparison of the PHQ-9 and ImPACT affective symptom cluster scores in measuring depression at baseline among college athletes. Archives of Clinical Neuropsychology, 35(5), 607.
13. Defreese, J., & Smith, A. (2014). Athlete social support, negative social interactions, and psychological health across a competitive sport season. Journal of Sports Sciences, 32(15), 1454-1464.
14. DiMillo, J., Hall, N. C., Ezer, H., Schwarzer, R., & Körner, A. (2017). The Berlin Social Support Scales: Validation of the Received Support Scale in a Canadian sample of patients affected by melanoma. Journal of Health Psychology, 24(13), 1785-1795.
15. Egan, S. J., Wade, T. D., & Shafran, R. (2011). Perfectionism as a transdiagnostic process: A clinical review. Clinical Psychology Review, 31, 203-212.
16. Faul, F., Erdfelder, E., Lang, A.G., & Buchner, A. G. (2007). *Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical Sciences. Behavior Research Methods, 39(2), 175-191.
17. Fortune, S., & Hetrick, S. (2022). Suicide risk assessments: Why are we still relying on these a decade after the evidence showed they perform poorly? Australian & New Zealand Journal of Psychiatry, 56(12), 1529-1534.
18. Franklin, J. C., Ribeiro, J. D., Fox, K. R., Bentley, K. H., Kleiman, E. M., Huang, X., …Nock, M. K. (2017). Risk factors for suicidal thoughts and behaviors: A meta-analysis of 50 years of research. Psychological Bulletin, 143(2), 187-232
19. Gorczynski, P., Miller-Aron, C., Moore, M. A., & Reardon, C. (2024). The epidemiology of mental health symptoms and disorders amongst elite athletes and the evolution of mental health literacy. Mental Health Considerations in Athletes, 43(1), 1-11.
20. Gupta, R., Agrawal, R. (2021). Are the concerns destroying mental health of college athletes? A Qualitative analysis portraying experiences amidst COVID-19 ambiguities. Analyses of Social Issues & Public Policy, 21(1), 621-639.
21. Hui, Z., Guo, K., Huang, W. B., Wu, J., Ma, X., Jia, S., & Xing, Z. (2023). Relationships between college students’ exercise motivation and mental health: Chain mediating effect of perceived social support and resilience. Social Behavior & Personality, 51(8), 1-10.
22. Kontos, A. P., Covassin, T., Elbin, R. J., & Parker, T. (2012). Depression and neurocognitive performance after concussion among male and female high school and collegiate athletes. Archives of Physical Medicine and Rehabilitation, 93(10), 1751-1756.
23. Kroenke, K., Spitzer, R. L., & Williams, J. B. (2011). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606-613.
24. LoGalbo, A., DaCosta, A., & Webbe, F. (2022). Comparison of the PHQ-9 and ImPAC symptom cluster scores in measuring depression among college athletes. Applied Neuropsychology: Adult, 29(4), 703-709.
25. MacDonald, H., & Neville, T. (2023). Promoting college students’ mindfulness, mental health, and self-compassion in the time of COVID-19: Feasibility and efficacy of an online, interactive mindfulness-based stress reduction randomized trial. Journal of College Student Psychotherapy, 37(3), 260-278.
26. Madan, A., Frueh, C. B., Allen, J. G., Eillis, T. E., Rufino, K. A., Oldham, J. M., & Fowler, C. (2016). Psychometric reevaluation of the Columbia – Suicide Severity Rating Scale: Findings from a prospective inpatient cohort of severely mentally ill adult. Journal of Clinical Psychiatry, 77(7), 867-873.
27. Maron, B. J., Haas, T. S., Murphy, C. J., Ahluwalia, A., & Rutten-Ramos, S. (2014). Incidence and causes of sudden death in the U.S. college athletes. Journal of American College of Cardiology, 63(16), 1636-1643.
28. Melhem, N. M., Porta, G., Oquendo, M. A., Zelazny, J., Keilp, J. G., Iyengar, S., … Brent, D. A. (2019). Severity and variability of depression symptoms predicting suicide attempt in high-risk individuals, JAMAPsychiatry, 76(6), 603-613.
29. Miller, K. E., & Hoffman, J. H. (2009). Mental well-being and sport-related identities in college students. Sociology of Sport Journal, 26(2), 335-356.
30. Moore, M. A. (2017). Stepping outside of their comfort zone: Perceptions of seeking psychosocial services amongst college athletes. Journal of Issues in Intercollegiate Athletics, Special Issue, 130-144.
31. Moore, M. A., & Abbe, A. (2021). The National Association of Intercollegiate Athletics substance use and abuse survey. Journal of Issues in Intercollegiate Athletics, 14, 95-114.
32. Moore, M. A., Gorczynski, P., & Miller-Aron, C. (2022). Mental health literacy in sport: The role of the social work profession. Social Work, 67(3), online. National Collegiate Athletic Association. (2022). NCAA student-athlete well-being study.
33. Olmedilla, A., Aguilar, J. M., Ramos, L. M., Trigueros, R., & Cantón, E. (2022). Perfectionism, mental health, and injuries in women footballers. Journal of Sport Psychology, 31(1), 49-56.
34. Orsolini, L., Latini, R., Pompili, M., Serafini, C., Volpe, U., Vellante, F. … De Bearardis, D. (2020). Understanding the complex of suicide in depression: From research to clinics. Psychiatry Investigation, 17(3), 207-2021.
35. Otsuki, T., Tomata, Y., Zhang, S., Tanji, F., Sugawara, Y., & Tsuji, I. (2019). The association between emotional and instrumental social support and risk of suicide death: A population-based cohort study. Journal of Psychiatric Research, 114, 141-146.
36. Posner, K., Brown, G. K., Stanley, B., Brent, D. A., Yershova, K. V., Oquendo, M. A., Currier, G. W., Melvin, G. A., Greenhill, L., Shen, S., & Mann, J. J. (2011). The Columbia – Suicide Severity Rating Scale: Initial validity and internal consistency findings from three multisite studies with adolescents and adults. The American Journal of Psychiatry, 168(12), 1266-1277.
37. Prinz, B., Dvořák, J., & Junge, A. (2016). Symptoms and risk factors of depression during and after the football career of elite female players. British Journal of Medicine and Sport: Open Sport & Exercise Medicine, 2(1), e000124.
38. Raedeke, T. D., & Smith, A. L. (2001). Coping with stressors in sport: A conceptual model. Journal of Sport & Exercise Psychology, 23, 528-541.
39. Rao, A. L., Asif, I. M., Drezner, J. A., Toresdahl, B. G., & Harmon, K. G. (2015). Suicide in National Collegiate Athletic Association (NCAA) athletes. Sports Health, 7(5), 452-457.
40. Ribeiro, J. D., Huang, X., Fox, K. R., & Franklin, J. C. (2018). Depression and hopelessness as risk factors for suicide ideation, attempts and death: Meta-analysis of longitudinal studies. British Journal of Psychiatry, 212(5), 279-286.
41. Rosenthal, S. R., Noel, J. K., Edwards, Z. C., Sammartino, C. J., & Swanberg, J. E. (2023). Risk factors for suicide ideation among Rhode Island college students. Rhode Island Medical Journal, 106(3), 42-47.
42. Rothschild-Checroune, E., Gravelle, F., Dawson, D., & Karlis, G. (2013). Balancing academic and athletic time management: A qualitative exploration of first year student athletes’ university football experiences. Society and Leisure, 35(2), 243-261.
43. Salehian, M. H., Gursoy, R., Aftabi, G. R., Sardarudi, M. T., & Ancari, Z. D. (2012). Comparison of depression between university athletes and non-athletes. European Journal of Experimental Biology, 2(4), 1008-1011.
44. Schmiedehaus, E., Snyder, E., Perrotte, J., Deason, R., Howard, K., & Cordaro, M. (2023). The ongoing mental health plight of depressed college students: Clinical recommendations and the importance of early screening and detection. Issues in Mental Health Nursing, 44(6), 562-570.
45. Schulz, U., & Schwarzer, R. (2004). Long-term effects of spousal support on coping with cancer after surgery. Journal of Social and Clinical Psychology, 23(5), 716-732.
Substance Abuse and Mental Health Services Administration. (2017). National Survey of Drug Use and Health Survey.
46. Substance Abuse and Mental Health Services Administration. (2021). National Survey of Drug Use and Health Survey.
47. Sullivan, M., Moore, M., Blom, L. C., & Slater, G. (2020). Relationship between social support and depressive symptoms in collegiate student athletes. Journal for the Study of Sports and Athletes in Education, 14(3), 192-209.
48. Syndergaard, S., Borger, J., Klenzak, S., Grello, A., & Adams, A. (2023). Implementation of Columbia – Suicide Severity Rating Scale (C-SSRS) as a universal suicide risk screening tool in high volume emergency department. Archives of Suicide Research, 27(2), 769-779.
49. Wolanin, A., Hong, E., Marks, D., Panchoo, K., & Gross, M. (2016). Prevalence of clinically elevated depressive symptoms in college athletes and differences by gender and sport.
British Journal of Sports Medicine, 50(3), 167–171.

2024-07-03T13:38:41-05:00July 5th, 2024|General, Research, Sport Education, Sports Studies and Sports Psychology|Comments Off on Navigating Darkness: College Athlete Suicide, Support Systems, and Shadows of Depression

Coaches’ Perspectives of the Influence of Safe Sport-Related Education

Authors: Anthony Battaglia1, Ph.D., Gretchen Kerr2, Ph.D., and Stephanie Buono2, Ph.D.

Corresponding Author:

Anthony Battaglia, Ph.D., CMPC 

Faculty of Kinesiology and Physical Education 

University of Toronto 

55 Harbord Street, ON, Canada, M5S 2W6 


Anthony Battaglia, Ph.D., CMPC is a Postdoctoral Fellow and lecturer in the Faculty of Kinesiology & Physical Education at the University of Toronto. His research interests focus on youth athletes’ sport experiences, relational dynamics in sport, athlete maltreatment, and strategies for advancing developmentally appropriate and safe sport.  

Gretchen Kerr, Ph.D. is a Full Professor and Dean of the Faculty of Kinesiology and Physical Education at the University of Toronto. She is also a co-Director of E-Alliance, the Canadian Gender Equity in Sport Research Hub.

Stephanie Buono, Ph.D. is a research associate in the Faculty of Kinesiology & Physical Education at the University of Toronto and an instructor in the Department of Applied Psychology & Human Development at the University of Toronto.

Coaches’ Perspectives of the Influence of Safe Sport-Related Education 


To combat growing concerns of sport being unsafe for athletes, compulsory safe sport education has been developed worldwide. Much of this education has focused on the role of the coach, largely due to their position of power, prevalence rates that highlight coaches as common perpetrators of harm, and their direct contact with athletes. However, there is a lack of research examining the impact of such education for coaching-related outcomes. The purpose of this study was to explore the influences of safe sport training on coaches’ knowledge and confidence, efficacy to support others, stress about athlete well-being, and stress related to safe sport issues. In an online survey, 1365 coaches reported completion of any of 12 possible safe sport training courses and their knowledge and confidence, efficacy to support others, stress about athlete well-being, and stress related to safe sport issues. Regression analyses indicated that completing any of the 12 safe sport-related training courses was related to perceived increased efficacy to support others. Completing a higher number of safe sport training courses was related to perceived increases in efficacy to support others and knowledge and confidence, but not stress related to safe sport or athlete well-being. All 12 courses were related to increased knowledge and confidence, and several courses were related to increased efficacy to support others and reduced safe sport stress, while one course was related to reduced stress about athlete-well-being. Future research is needed to examine whether improvements in coaching outcomes associated with safe sport training translate into practice.

Key Words: Safe Sport; Coaches; Education; Coaching Outcomes;

Over the last several years, numerous reports of concerning behaviors in sport, such as maltreatment have emerged worldwide (15, 25). Maltreatment, which refers to “volitional acts that result in or have the potential to result in physical injuries and/or psychological harm” (12, p. 3), which include psychological, sexual, physical abuse, and neglect, harassment, bullying, and discrimination. To combat such concerns, policies and educational initiatives have been developed and implemented under the term ‘safe sport’ (26). The term safe sport initially emerged in response to scandals involving sexual abuse but has since expanded to refer to participation in sport free from all forms of violence, abuse, discrimination, and harassment (21, 39). More recently, broader conceptualizations of safe sport have also considered issues of environmental and physical safety (e.g., dysfunctional equipment, performance enhancing drugs), and the optimization of the sport experience (i.e., inclusive, accessible, growth-enhancing, and rights-based participation for all) (18). To advance safe sport, compulsory education has been developed; examples of existing safe sport education programmes include Australia’s Play by the Rules, U.S. Center for SafeSport Training, and the UK’s Child Protection in Sport Unit (24, 26).

Although safe sport education is needed for all sport stakeholders, including athletes, coaches, parents, administrators, officials and support staff, to-date, education has focused largely on coach-athlete dynamics, addressing issues such as harmful coaching practices, power relations, and duty to report harm (24, 26). There is a strong rationale for safe sport training focused on coaches. Consistent across many bodies of research in sport is acknowledgement of the presence and effects of the position of power and authority held by coaches over stakeholders in the sport ecosystem, including subordinate coaches, parents, athletes, and administrators (23, 38). When used inappropriately, these positions of power leave others vulnerable to experiences of harm. For example, psychological abuse (or what some refer to as psychological violence), the most prevalent form of athlete maltreatment, is most often perpetrated by coaches (42, 45, 48). Given their direct contact with other coaches, support staff, athletes and/or teams daily, coaches also significantly impact the type of culture promoted (e.g., win-at-all-costs versus caring or athlete-centred) and the nature and quality of athletes’ experiences (32). Coaches who are provided professional development and educational opportunities regarding positive sport practices are more likely to create environments where athletes experience enjoyment, competence, meaningful relationships, learning, satisfaction, reduced anxiety, and sport maintenance (6, 16, 36).

Although growing awareness of athlete maltreatment and the role of the coach in preventing these experiences has resulted in the proliferation of safe sport education initiatives for coaches globally, little research exists on the impact of such education for coaching-related outcomes (24, 26). In 2013, McMahon (28) investigated how a narrative pedagogical approach (i.e., athletes’ stories) might help swim coaches from amateur and elite levels understand the welfare implications for athletes subjected to emotionally or physically abusive coaching practices. Findings revealed that coaches gained increased empathy and undertook a more athlete-centered approach to coaching post-education, however, dominant cultural ideologies (e.g., winning) persisted in the coaches’ thinking and practice. Likewise, in 2018, Nurse (30) examined child sexual abuse prevention training for adults who work with children in schools, churches, and athletic leagues; with regards to coaches specifically, the training improved coaches’ knowledge on the topic and increased their confidence in their ability to identify abuse. These preliminary findings highlight the potential benefits of training for coaches; however, it is important to note that the education programmes were restricted to specific populations, sports, forms of harm, small sample sizes, and the effects of long-term behavioral change remained unclear. Further research examining the impact of safe sport training for coaches is required.

In Canada, the country of interest in this study, safe sport educational modules (e.g., NCCP Make Ethical Decisions, Safe Sport Training) (7, 9) have been developed by the Coaching Association of Canada (CAC), which is responsible for certifying and educating coaches across Canada. The CAC has also promoted safe sport standards and expectations for organizations and its coaches, including the Responsible Coaching Movement- a pledge to learn and apply consistent safety principles. The pillars of the Responsible Coaching Movement include the Rule of Two, which attempts to ensure all interactions and communications are in open, observable, and justifiable settings; background screening; and ethics training (8). In the province of Ontario, the Coaches Association of Ontario- an independent, non-profit organization that supports coaches from community level to high performance across all sports in Ontario- has adopted similar safe sport efforts and developed resources, such as Safe Sport 101 and the Ontario Coaches Conference (10). The goals of such initiatives include but are not limited to improving the knowledge of coaches with respect to safe sport, increasing their confidence in enacting desirable coaching behaviors, creating positive sport climates, and facilitating the holistic development of athletes. To-date, the extent to which these educational initiatives meet these goals for Canadian coaches has not been examined.

While safe sport education for coaches has commonly focused on enhancing knowledge of harmful or prohibited conduct, enhancing confidence in using desired behaviors, and supporting stakeholders’ (e.g., athletes, coaches, support staff) development and well-being, there remains a lack of research examining the influence of safe sport training on coaching-related outcomes (24, 26). In this study, the constructs of knowledge, confidence, efficacy, and stress were of interest. Despite recognizing their influential role, many coaches admit inadequate knowledge to cultivate safe sport environments (25); as cultivating safe sport environments is also a collective effort, it remains important that coaches feel efficacious in their ability to support all participants (31). Given the prevalence of mental health challenges in sport, coaches have expressed stress related to supporting athletes’ mental well-being (1, 3). Further, in response to the public attention paid to cases of athlete maltreatment and the focus on coaches as perpetrators of harm, coaches have reportedly felt stress about potential false accusations; specifically, concerns of negative touch have been identified in research and practice, resulting in coaches and sport personnel being fearful and unsure of how to be around athletes with whom they interact (40).

The purpose of this study therefore to explore the influences of safe sport training on Ontario coaches’ knowledge and confidence, efficacy to support others, stress about athlete well-being, and stress related to safe sport issues. Specifically, the first objective was to examine whether safe sport training improved coaches’ knowledge and confidence, efficacy to support others, stress about athlete well-being, and stress related to safe sport issues. The second objective was to examine whether the effect of safe sport training on coaches increased with the number of safe sport training courses. The third objective was to examine whether certain courses were related to coaches’ knowledge and confidence, efficacy to support others, stress about athlete well-being, and stress related to safe sport issues.



This study was conducted in partnership with the Coaches Association of Ontario (CAO). CAO is an independent, non-profit organization that supports coaches across all levels and sports in Ontario. Ontario has the largest population of all provinces in Ontario with over 15 million people and one in four Ontarians have coached in their lifetime (10). The CAO selected the safe sport-related courses of interest for evaluation (see Table 1). As such, within the context of the current study, a broad perspective of safe sport (i.e., from injuries to drug-free sport, planning appropriate practices, and maltreatment) was adopted. Upon receiving approval from the University of Toronto Health Sciences Research Ethics Board, coaches were contacted through the Coaches Association of Ontario (CAO) email listserv and social media posts (Facebook, Instagram, Twitter). Recruitment communication provided details about study eligibility/requirements, the purpose of the study, the voluntary nature of the study, confidentiality and anonymity, and the link to the online survey. The survey was administered with RED Cap electronic data capture. Participants were required to meet the following eligibility criteria to complete the online survey: 1) Ontario resident; 2) over the age of 16; and 3) had coached in the last two years. Following the confirmation of eligibility, participants were able to complete the survey, which took approximately 15-25 minutes (M=19.25) to complete.

Table 1. An overview of the Safe Sport Education modules evaluated in the current study.

NCCP Emergency Action Planning completion of this module, coaches will be able to: describe the importance of having an EAP; identify when to activate the EAP; explain the responsibilities of the charge person and call person when the EAP is activated; and create a detailed EAP that includes all required information for responding to an emergency.
NCCP Planning a Practice completion of this module, coaches will be able to: explain the importance of logistics in the development of a practice plan; establish an appropriate structure for a practice; and identify appropriate activities for each part of the practice. To receive full credit for this module, coaches must also complete NCCP Emergency Action Planning.
NCCP Making Head Way completion of this module, coaches will understand how to: prevent concussions; recognize the signs and symptoms of a concussion; what to do when they suspect an athlete has a concussion; and ensure athletes return to play safely.
NCCP Leading Drug-Free Sport completion of this module, coaches will be able to: understand and demonstrate their role in promoting drug free sport; assist athletes to recognize banned substances and the consequences as identified by the Canadian Centre for Ethics in Sport; reinforce the importance of fair play and the NCCP Code of Ethics; educate and provide support to athletes in drug testing protocols; and inform athletes on nutritional supplements.
NCCP Prevention and Recovery completion of this module, coaches will be able to: identify common injuries in sport, prevention and recovery strategies; design and implement programs/activities to optimize athlete training, performance and recovery; and support athletes’ return to sport through awareness and proactive leadership.
Commit to Kids completion of this module, coaches will be able to: understand and recognize child sexual abuse and the grooming process; ways in which to handle disclosures of sexual abuse; the implications of sexual abuse; how to create a child protection code of conduct; and ways in which to enhance child and youth safety in sport.
Standard First Aid and CPR completion of this module, coaches will be able to: understand and apply vital life-saving knowledge/skills essential for meeting a variety of workplace/professional requirements.
HeadStartPro completion of this module coaches will be able to: understand and develop a set of coaching tools to improve team performance and injury-prevention; and assist athletes and/or teams in achieving their full potential with performance-driven injury prevention training.
NCCP Making Ethical Decisions completion of this module coaches will be to: analyze challenging situations and determine the moral, legal, or ethical implications; and apply the NCCP Ethical Decision-Making Model to respond in ways that are consistent with NCCP Code of Ethics.
NCCP Empower+ (Creating Positive Sport Environments) completion of this module, coaches will be able to: describe the characteristics and benefits of participant-centered coaching; explain the types of harm that may occur when a coach misuses their power; respond to suspicions or knowledge of maltreatment; and implement positive coaching strategies to foster learning, performance, and create a safe sport environment.
CAC Safe Sport completion of this module, coaches will be able to: understand the critical role of all stakeholders in promoting safe sport, how the misuse of power leads to maltreatment, and principles of the Universal Code of Conduct; understand types of maltreatment and how to recognize signs and symptoms; and respond when maltreatment is suspected and create a safe sport culture for all participants.
Respect in Sport completion of this module, coaches will be able to: recognize, understand, and respond to issues of bullying, abuse, harassment, and discrimination.

Note. For further detail on course descriptions and/or objectives see the corresponding webpages indicated in the table.


Participants were 1365 coaches from the Coaches Association of Ontario (CAO). Of the respondents, 61% identified as men (n=823), 38% identified as women (n=514; n=28 did not disclose), 86% identified as White (n=1087), while 4% (n=53) identified as Black, 4% (n=51) identified as East/Southeast Asian, 2% (n=31) identified as Indigenous, and less than 2% identified as Latinx (n=19), South Asian (n=18), Middle Eastern (n=16), or another race category (n=27). Coaches reported working in a variety of contexts including grassroots (e.g., recreational, community sport, house league, intramural; n=273, 22%), school sports (e.g., primary and secondary school; n=141, 11%), development (e.g. competitive, club, travel, city, all-star; n=600, 49%), post-secondary (e.g., Support, CCAA, OUA, Inter-university; n=74, 6%), provincial (e.g., Canada Games, National Championships, OHL; n=90, 7%), international (e.g., International Competitions, Worlds, Pan Am, Commonwealth, Olympics; n=36, 3%), and masters or professional (e.g., Senior, NHL, NBA, CEBL; n=20, 2%). Coaches’ tenure in their current position ranged from 1-10 years (n=804, 65%), 11-20 years (n=238, 19%), and more than 20 years (n=194, 16%). Training in safe sport was required for 78% of coaches (n=782) and provided free of cost for 51% of coaches (n=535).


Safe sport training was measured with a “yes” or “no” response from coaches to indicate whether they had taken each of the following courses: NCCP[1] Emergency Action Planning, NCCP Planning a Practice, NCCP Making Head Way, NCCP Leading Drug Free Sport, NCCP Prevention and Recovery of Injury, Commit to Kids, Standard First Aid and CPR, HeadStart, NCCP Make Ethical Decisions, NCCP Empower+ (Creating Positive Sport Environments), CAC Safe Sport Training, Respect in Sport.

Knowledge & confidence to create a safe sport environment was measured using a 3-item scale (a=.7), which asked coaches about their knowledge of safe sport concepts and their confidence in creating a safe sport environment. Example items included, “I am confident in my abilities to create a safe sport environment” and “I am familiar with the responsible coaching movement.” Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Safe sport stress was measured using a 3-item scale (a=.68), which asked coaches about the stress they experience over creating a safe sport environment. An example item includes, “I often stress about being the subject of a harassment or abuse claims”. Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Stress about athlete well-being was measured with 2 items (a=.59): “I often stress about my ability to manage athletes’ mental well-being”, and “I often stress about my ability to manage athletes’ physical well-being.” Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Efficacy to support others was measured using a 5-item scale (a=.87), which asked coaches about how confident they feel in their ability to support athletes and other coaches. An example item includes “I am confident in my abilities to support my athletes with performance issues”. Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

[1] NCCP refers to the National Coaching Certification Program offered by the Coaching Association of Canada.

Safe sport stress was measured using a 3-item scale (a=.68), which asked coaches about the stress they experience over creating a safe sport environment. An example item includes, “I often stress about being the subject of a harassment or abuse claims”. Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Stress about athlete well-being was measured with 2 items (a=.59): “I often stress about my ability to manage athletes’ mental well-being”, and “I often stress about my ability to manage athletes’ physical well-being.” Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Efficacy to support others was measured using a 5-item scale (a=.87), which asked coaches about how confident they feel in their ability to support athletes and other coaches. An example item includes “I am confident in my abilities to support my athletes with performance issues”. Coaches responded to each item on a scale from 1 (strongly disagree) to 5 (strongly agree).

Data Analysis

To investigate the first research objective, an initial correlation analysis was conducted to examine whether having any safe sport training was related to increases in coaching outcomes. The safe sport training variable was transformed so that coaches who answered “yes” to completing any of the safe sport training courses were coded as 1 and coaches who had answered “no” to completing all the safe sport training courses were coded as 0 (i.e., no SS training=0, any SS training=1). This variable was included in a correlation analysis with all coaching outcomes: knowledge & confidence, safe sport stress, stress over athlete well-being, and efficacy to support others. To investigate the second research objective, four separate linear regression models were constructed with the sum of completed safe sport training courses (range =1-12) as the independent variable, and the following coaching outcomes as respective dependent variables: knowledge & confidence, safe sport stress, stress about athlete well-being, and efficacy to support others. In all four models, the coaching context, whether training was required (0=no, 1=yes), and whether training was free (0=no, 1=yes) were included as covariates. To address the third research objective, ANOVAs were conducted with individual safe sport courses as independent variables, and the following coaching outcomes as dependent variables: knowledge & confidence, efficacy to support others, safe sport stress, stress about athlete well-being and efficacy to support others. All analyses were conducted using IBM SPSS Statistics (Version 28) (20).


Safe Sport Training & Coaching Outcomes

Range, mean, and standard deviation scores for all variables included in subsequent analyses are included in Table 2. Of the coaches in this sample, 65% (n=890) reported completing at least one of the education courses, while 35% (n=475) reported not having taken any of the education courses. Results of the correlation analysis (Table 3) demonstrate that having any safe sport training was significantly related to increases in efficacy to support others, but not knowledge and confidence, safe sport stress, or stress about athlete well-being.

Table 2. Descriptive statistics for all variables

Coaching Context (0=Grassroots)0-71.811.37
Training Required (0=No)0-1.59.49
Training Free (0=No)0-1.49.50
Any Safe Sport Training0-1.6.13
Number of Safe Sport Training0-123.643.42
Knowledge & Confidence-4-201
Safe sport stress-4-201
Stress over athlete well-being-4-201
Efficacy to Support-4-201
Table 3. Correlations between any safe sport training and coaching outcomes
Any Safe Sport TrainingKnowledge ConfidenceSafe Sport StressAthlete WB StressEfficacy to Support
Any Safe Sport Training1.00.06*.04.002-.03
Knowledge Confidence.06*1.00-.02.00.29**
Safe Sport Stress.04-.021.00.34**-.09**
Athlete WB Stress.002.00.34**1.00-.20**
Efficacy to Support-.03.29**-.09**-.20**1.00
**. Correlation is significant at the 0.01 level
*. Correlation is significant at the 0.05 level

Number of Safe Sport Training & Coaching Outcomes

Figure 1 demonstrates the number of safe sport courses taken by coaches in this sample based on influential covariates such as coaching context, training requirement, and training accessibility (i.e., whether the training was provided free of cost). Significantly more safe sport courses were completed by coaches in Post-Secondary, Provincial, International, Masters and Professional contexts, and by coaches for whom training and education is required and free. 

Initial correlation analysis (Table 4) demonstrated being a coach at a high level of competition (e.g., provincial, international) was related to taking more safe sport courses, higher knowledge and confidence, and higher efficacy to support others. Having access to free training was related to taking more safe sport courses and higher knowledge and confidence. Finally, taking more safe sport training courses was related to higher knowledge and confidence and efficacy to support others. Safe sport stress and stress about athlete well-being were not related to any of the independent variables.

Table 4. Correlations between number of safe sport training courses, covariates and outcome variables
Coaching ContextTraining RequiredTraining FreeSafe Sport TrainingKnowledge ConfidenceSafe Sport StressAthlete WB StressEfficacy to Support
Coaching Context1.00-.04-.03.11**.07**.01.00.08**
Training Required-.041.00.11**-.02.08**.06.03-.05
Training Free-.03.11**1.00.09**.08*.00-.06.01
Safe Sport Training.11**-.02.09**1.00.26**.05.01.10**
Knowledge Confidence.07**.08**.08*.26**1.00-.02.00.29**
Safe Sport Stress.**-.09**
Athlete WB Stress.00.03-.**1.00-.20**
Efficacy to Support.08**-.05.01.10**.29**-.09**-.20**1.00
**. Correlation is significant at the 0.01 level
*. Correlation is significant at the 0.05 level

The results of the first regression analysis (Table 5) demonstrated that the number of safe sport training courses coaches completed was related to increases in knowledge and confidence and efficacy to support others, when training requirements, access to training, and context of the sport environment were held constant. The number of safe sport training courses coaches took was not related to safe sport stress or athlete well-being stress.

Table 5. Linear Regression Analyses for General Coach Training
Knowledge & ConfidenceSafe Sport StressAthlete WB StressEfficacy to Support
Coaching Context.*.02
Training Required.09*.
Training Free.08*.
Safe Sport Training.31**.**.01
Adj. R-Square. 
**Coefficient is significant at the 0.01 level
*Coefficient is significant at the 0.05 level

Individual Safe Sport Courses and Coaching Outcomes

The results of the final analysis demonstrated that all courses were significantly related to improved knowledge and confidence. NCCP Emergency Action Planning, NCCP Leading Drug Free Sport, Commit to Kids, HeadStartPRO, and NCCP Empower+ (Creating Positive Sport Environments) were significantly related to reduced safe sport stress. Commit to Kids was significantly related to reduced athlete well-being stress. Finally, NCCP Planning a Practice, NCCP Leading Drug-free Sport, NCCP Prevention and Recovery, Commit to Kids, HeadStartPRO, NCCP Empower+ (Creating Positive Sport Environments), and CAC Safe Sport were significantly related to efficacy to support others (Table 6).

Table 6. Effects of Individual Safe Sport Courses
Knowledge ConfidenceSafe Sport StressAthlete WB StressEfficacy to Support Others
NCCP Emergency Action Planning60.97<.0015.67.0171.45.2293.75.053
NCCP Planning a Practice53.82<.001.13.722.44.5097.23.007
NCCP Making Head Way64.15<.001.10.754.08.772.35.557
NCCP Leading Drug-free Sport72.82<.0015.<.001
NCCP Prevention and Recovery47.18<.0013.<.001
Commit to Kids35.88<.0015.16.0238.91.00311.29<.001
Standard First Aid and CPR17.96<.001.31.580.69.4069.73.002
NCCP Making Ethical Decisions22.26<.001.17.680.01.931.01.91
NCCP Empower+ (Creating Positive Sport Environments)15.21<.0017.92.04.315.57516.42<.001
CAC Safe Sport89.17<.001.16.6903.91.5328.41.004
Respect in Sport32.62<.001.07.797.07.7973.64.056


The purpose of this study was to explore the influences of safe sport training on sport coaches’ knowledge and confidence, safe sport-related stress, efficacy to support others, and stress about athlete well-being. Specific focus was directed towards examining the relationship between the number of safe sport courses completed and the effects of specific safe sport courses for these coaching outcomes. The results of this study demonstrated that having any training or education was related to increased efficacy to support others. Having completed a higher number of safe sport training courses was related to increased efficacy to support others and knowledge and confidence, and all safe sport courses were related to increased knowledge and confidence.  

Although a plethora of safe sport education exists to-date, a prominent criticism has been the lack of empirical evaluations examining the impact or effectiveness of such training (24, 26). The findings of the current study help to address this knowledge gap by providing preliminary, empirical evidence regarding the influence of safe sport education. According to the results, coaches in more professional contexts took more safe sport training courses, which supports the notion that at elite levels of sport, coaches may have more access to professional development opportunities and/or devote more time improving their coaching skills (11, 27). Coaches who were provided access to free training in the current study also took more safe sport courses. These findings suggest that when provided the opportunity, coaches engage in professional development, however, as issues of cost and accessibility remain prevalent barriers, the advancement and development for many coaches remains limited (19, 43. Online modalities have been advocated as a cost-effective, time efficient, and readily accessible way to educate coaches (13, 14) yet, for many coaches, online professional development opportunities still present financial demands. For example, of the twelve courses examined in the current study, only three (e.g., NCCP Emergency Action Planning, CAC Safe Sport, NCCP Making Headway) are listed as online and free for coaches; in the current study, it was not known if affiliated organizations where coaches instruct reimbursed education/training and, if so, for which courses. Access or lack thereof to safe sport-related education may impact the extent to which safe, inclusive, and welcoming spaces are promoted by all coaches (22, 47). This is particularly important for coaching at the youth sport level where the delivery of sport programmes is highly dependent on volunteers who, despite recognizing their critical role for nurturing developmentally appropriate and safe environments, often lack the requisite knowledge to do so (2, 44, 46).

The completion of more safe sport training courses and all courses examined in the current study was related to enhanced coaches’ knowledge and confidence. Exposing coaches to diverse topics which include but are not limited to safety, positive development, harmful practices, and mental health, are critical to improving coaches’ awareness and ability to create safe sport environments (6, 28, 30). The coaches also reported increased knowledge of the Rule of Two and the Responsible Coaching Movement; these safe sport efforts provide additional safety principles for Ontario and Canadian coaches more broadly on background screening, appropriate interactions, and ethics training (8). Findings may be interpreted to suggest that not only does safe sport education positively influence coaches’ knowledge and confidence to create safe environments but also facilitates greater awareness of safe sport efforts in the Canadian sport context, thus providing coaches with a more comprehensive perspective on ways to stimulate safer sport.

Nurturing athletes’ holistic development is a key responsibility of coaches; however, coaches may not have the necessary education and training to adequately support their athletes (41). The current findings indicate that the completion of more safe sport education as well as specific courses (e.g., NCCP Empower+, CAC Safe Sport) may nurture coaches’ expertise and confidence to actively support their athletes with personal and performance challenges. The extent to which athletes report positive coach-athlete dynamics and feel supported in their relationships with coaches has been known to influence whether they experience learning, growth, and safe sport environments (32). Safe sport training also influenced coaches’ confidence to support coaching peers/support staff with personal and performance issues; these findings are particularly important as learning by doing, having a coach mentor, and observing others are important sources of knowledge and development for coaches (43). Collectively, the improvements in coaches’ efficacy to support others (athletes and coaches) suggests that safe sport training may serve as an effective mechanism through which knowledge dissemination and learning amongst stakeholders is achieved.

Many coaches (uninformed on the benefits of positive touch) have adopted a risk-averse perspective when interacting with athletes (i.e., “no touching”) to avoid being accused of misconduct or having their behaviors misconstrued as harmful (33, 34). In the current study, no significant relationship resulted between the number of safe sport training courses completed and coaches’ perceived safe sport stress (e.g., fear of maltreatment allegations). Specific courses were identified as decreasing safe sport stress, however, some of the courses (e.g., NCCP Emergency Action Planning, HeadStartPro, NCCP Leading Drug-free Sport) focus on physical injury prevention and/or drug-free sport and do not necessarily provide broader content on maltreatment that might warrant the reported lower coach stress regarding potential accusations of harm or safe sport issues. Although coaches have commonly reported concerns about touching in sport (33), there has also been growing awareness of psychological harm and toxic cultures in sport (38, 48). The lack of reported stress regarding safe sport concerns may be reflective of coaches being less fearful of false accusations related to psychological forms of harm as opposed to sexual harms. As the survey questions referred to coach stress in relation to abuse and harassment claims broadly, further research attention is needed to assess whether education may impact coaches’ safe sport stress differently depending on the form of harm (e.g., sexual versus psychological).

It is also interesting that while safe sport education was related to coaches’ improved efficacy to support athletes with personal and performance issues, the number of completed courses was not significantly related to stress about managing athlete physical and mental well-being. Only one course (Commit to Kids) reduced coaches’ perceived stress for managing athlete well-being. Commit to Kids focuses exclusively on providing education on sexual harms; while education on sexual harms is needed to advance safe sport, psychological harm and neglect are reported far more frequently by athletes (25, 48) and thus coaches’ perceptions of their ability to manage athletes’ well-being may be limited in scope.

            NCCP Empower+ (Creating Positive Sport Environments) was associated with enhanced knowledge and confidence, improved efficacy to support others, and lower safe sport stress, whereas CAC Safe Sport Training was linked to improved knowledge and confidence and efficacy to support others. Interestingly, Commit to Kids was the only course to positively impact all coaching outcomes, despite focusing exclusively on sexual harms. As sexual harm continues to receive the most media and research attention (4, 25), education on sexual harms may be interpreted by coaches and those in the sport community to be most relevant and important for creating safe sport. Further, in Ontario and Canada more broadly, sport organizations frequently identify course equivalents where coaches may complete different courses, including CAC Safe Sport Training, Respect in Sport, NCCP Empower+, and Commit to Kids but still satisfy the safe sport-related requirements needed to instruct. The lack of an integrated approach and the various safe sport education options available may expose coaches to different experiences and levels of learning, thus providing a plausible explanation for the reported influences on coaching outcomes in the current study. To advance safe sport,evidence-informed education for coaches and stakeholders more broadly is needed (5, 47).

Limitations and Future Directions

Although this study contributes to research and practice in safe sport by providing insights into the reported benefits of safe sport education for coaches, the findings must be interpreted within the context of the current study. Considering the CAO selected the safe sport-related courses of interest for evaluation, a broad perspective of safe sport (i.e., injuries, drug-free sport, planning appropriate practices, maltreatment) was required. The data were also collected from coaches in a specific geographic region (Ontario, Canada) and thus many of the safe sport courses evaluated were exclusive to this coaching sample. The courses evaluated in the current study should not be considered an exhaustive list of all safe-sport courses; for example, since the completion of the study, several courses (e.g., Support Through Sport, Safe Sport 101 Playbook) have been revised and/or developed. Additionally, as the sport domain has been referred to one that reinforces toxic cultures, there are several education courses in Ontario and Canada more broadly on creating positive culture and inclusive environments (e.g., NCCP Coaching Athletes with a Disability), that were not included and require future consideration regarding their impact on coaches and advancing safe sport. 

The study findings highlighted a relationship between safe sport education and improvements in coach knowledge and confidence and efficacy to support others, suggesting that practitioners should explore ways to make safe sport education free of cost and accessible. However, as this study did not assess knowledge translation, future research is needed to examine if coaches’ improved knowledge, confidence and efficacy from education contributes to behavior change and the use of more developmentally appropriate and safe coaching practices. Organizational influence also remains an area of interest; for example, it would be beneficial to explore how an organization’s cultural values, priorities (e.g., win-at-all-costs vs holistic development), and support (e.g., free training), may impact coach education uptake and subsequently the effectiveness of safe sport education on coaching outcomes. Future researchers may consider a case study approach to examine the impact of safe sport education for coaches within a specific organization; for example, Likert-scales may be used to assess attitudes, beliefs, and perceptions, semi-structured interviews may help to gain deeper insights on coaches’ interpretations regarding safe sport courses, and participant observation may shed light on issues of coach behavior change resulting from safe sport education.


Safe sport education for coaches has been consistently advocated as a recommendation for advancing safe, inclusive, and welcoming environments, however, the influence of safe sport education remains largely unknown (24, 26). The current study contributes to the sport literature by providing an examination of the influences of safe sport training for coaches. Findings revealed a relationship between the number of safe sport training courses coaches completed and increases in their knowledge and confidence and efficacy to support others. However, the number of safe sport training courses completed was not associated with stress related to safe sport matters or athlete well-being. All safe sport courses were reportedly associated with improved coach knowledge and confidence. Several training courses were also linked to improvements in coaches’ efficacy to support others and reductions in their safe sport stress, with only one course contributing to coaches’ reduced stress related to athlete-well-being. Although the findings suggest favorable influences of safe sport training for coaches, the current study did not assess behavioral change. Future research is needed to explore whether reported improvements (e.g., knowledge and confidence) associated with safe sport education translates to coaching practice.

Applications in Sport

Safe sport education in the current study was reportedly associated with enhanced coach knowledge and confidence to create safe environments and efficacy to support athletes and other coaches/support staff. Unfortunately, as a large portion of the sport sector is run by a volunteer workforce (e.g., volunteer coaches), sport organizations remain reluctant to enforce training requirements for fear of further burdening these coaches who frequently report stress and burnout (2, 35). However, the extent to which sport organizations and their leaders prioritize and support safe sport, has been shown to impact the effectiveness of safe sport efforts (17, 37, 49). In some cases, merely having safe sport education initiatives may have little impact on creating and sustaining safer environments and appear as superficial gestures towards change, further reproducing harms (29, 31). Sport and coaching organizations are confronted with the challenge of maintaining low time and cost demands for many volunteer coaches while also providing adequate education for volunteer (and paid) coaches (19, 46).


The authors would like to thank the coaches who participated in this study along with Coaches Association of Ontario who contributed to the design and recruitment of this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.


  1. Battaglia, A., & Kerr, G. (2022). Examining the impact of COVID-19 on sport coaches. International Sport Coaching Journal10(1), 102-111.
  2. Baxter, H., & Misener, K. E. (2022). Retaining volunteer coaches in child and youth sport. In Routledge Handbook of Coaching Children in Sport (pp. 412-420). Routledge.
  3. Bissett, J. E., Kroshus, E., & Hebard, S. (2020). Determining the role of sport coaches in promoting athlete mental health: A narrative review and Delphi approach. BMJ Open Sport & Exercise Medicine6(1), e000676.
  4. Bjørnseth, I., & Szabo, A. (2018). Sexual violence against children in sports and exercise: A systematic literature review. Journal of child sexual abuse27(4), 365-385.
  5. Brackenridge, C. H., & Rhind, D. (2014). Child protection in sport: reflections on thirty years of science and activism. Social Sciences3(3), 326-340.
  6. Callary, B. & Gearity, B. (2019) Coach education and development in sport: Instructional strategies. Routledge.
  7. Coaching Association of Canada (2023). NCCP make ethical decisions.
  8. Coaching Association of Canada (2023). Responsible coaching movement.
  9. Coaching Association of Canada (2023). Safe Sport training.
  10. Coaches Association of Ontario (2023). Programs and resources.
  11. Côté, J., Erickson, K., & Duffy, P. (2013). Developing the expert performance coach. In D. Farrow, J. Baker, & C. MacMahon (Eds.), Developing sporting expertise (2nd ed., pp. 96-112). Routledge.
  12. Crooks, C. V. & Wolfe D.A. (2013). Child abuse and neglect. In E.J.  Mash & R. A. Barkley (Eds.), Assessment of Childhood Disorders (pp. 1-17). Guilford Press.
  13. Cushion, C. J., & Townsend, R. C. (2019). Technology-enhanced learning in coaching: A review of literature. Educational Review71(5), 631-649.
  14. Driska, A., & Nalepa, J. (2020). Self-paced online learning to develop novice, entry-level, and volunteer coaches. In B. Callary & B. Gearity (Eds.), Coach education and development in sport: Instructional strategies (pp. 166–177). Routledge.
  15. Fortier, K., Parent, S., & Lessard, G. (2020). Child maltreatment in sport: Smashing the wall of silence: a narrative review of physical, sexual, psychological abuses and neglect. British Journal of Sports Medicine54(1), 4-7.
  16. Gould, D. (2013). Effective education and development of youth sport coaches. President’s Council on Fitness, Sports and Nutrition: Research Digest14(4), 1-10.
  17. Gurgis, J. J., & Kerr, G. A. (2021). Sport administrators’ perspectives on advancing safe sport. Frontiers in sports and active living3, 630071.
  18. Gurgis, J. J., Kerr, G., & Battaglia, A. (2023). Exploring stakeholders’ interpretations of safe sport. Journal of Sport and Social issues47(1), 75-97.
  19. Gurgis, J. J., Kerr, G. A., & Stirling, A. E. (2020). Investigating the barriers and facilitators to achieving coaching certification. International Sport Coaching Journal, 7(2), 189-199.
  20. IBM Corp. (2022). IBM SPSS Statistics for MacOs (Version 28.0). IBM Corp. International Olympic Committee (2021). IOC Safe Sport initiatives: Overview.
  21. International Olympic Committee (2021). IOC Safe Sport initiatives: Overview.
  22. Johnson, N., Hanna, K., Novak, J., & Giardino, A. P. (2020). US center for SafeSport: Preventing abuse in sports. Women in Sport and Physical Activity Journal28(1), 66-71.
  23. Jowett, S., & Wachsmuth, S (2020). Power in coach-athlete relationships: The case of the women’s artistic gymnastics. In G. Kerr (Ed.), Women’s artistic gymnastics: Sociocultural perspectives (pp. 121-142). Routledge.
  24. Kerr, G., Stirling, A., & MacPherson, E. (2014). A critical examination of child protection initiatives in sport contexts. Social Sciences3(4), 742-757.
  25. Lang, M. (2021). Routledge handbook of athlete welfare. Routledge.
  26. MacPherson, E., Battaglia, A., Kerr, G., Wensel, S., McGee, S., Milne, A., … & Willson, E. (2022). Evaluation of publicly accessible child protection in sport education and reporting initiatives. Social Sciences11(7), 310.
  27. Martens, R. (2018). Successful coaching. Human Kinetics.
  28. McMahon, J. (2013). The use of narrative in coach education: The effect on short-and long-term practice. Sports Coaching Review2(1), 33-48.
  29. Nite, C., & Nauright, J. (2020). Examining institutional work that perpetuates abuse in sport organizations. Sport Management Review23(1), 117-129.
  30. Nurse, A. M. (2018). Coaches and child sexual abuse prevention training: Impact on knowledge, confidence, and behavior. Children and Youth Services Review88, 395-400.
  31. Owusu-Sekyere, F., Rhind, D. J., & Hills, L. (2022). Safeguarding culture: towards a new approach to preventing child maltreatment in sport. Sport Management Review25(2), 300-322.
  32. Pills, S (2018). Perspectives on athlete-centred coaching. Routledge.
  33. Piper, H. (2014). Fear, risk, and child protection in sport: Critique and resistance. In H. Piper (Ed.), Touch in Sports Coaching and Physical Education (pp. 167-186). Routledge.
  34. Piper, H., Taylor, B., & Garratt, D. (2012). Sports coaching in risk society: No touch! No trust! Sport, Education and Society17(3), 331-345.
  35. Potts, A. J., Didymus, F. F., & Kaiseler, M. (2019). Exploring stressors and coping among volunteer, part-time and full-time sports coaches. Qualitative Research in Sport, Exercise and Health11(1), 46-68.
  36. Reynders, B., Vansteenkiste, M., Van Puyenbroeck, S., Aelterman, N., De Backer, M., Delrue, J., … & Broek, G. V. (2019). Coaching the coach: Intervention effects on need-supportive coaching behavior and athlete motivation and engagement. Psychology of Sport and Exercise43, 288-300.
  37. Rhind, D. J., & Owusu-Sekyere, F. (2020). Evaluating the impacts of working towards the International Safeguards for Children in Sport. Sport Management Review23(1), 104-116.
  38. Roberts, V., Sojo, V., & Grant, F. (2020). Organisational factors and non-accidental violence in sport: A systematic review. Sport Management Review23(1), 8-27.
  39. Safe Sport International (2021). Abuse of athletes happens.
  40. Tam, A., Kerr, G., & Stirling, A. (2020). Influence of the# MeToo movement on coaches’ practices and relations with athletes. International sport coaching journal8(1), 1-12.
  41. Thelwell, R., Harwood, C., & Greenlees, I. (2017). The psychology of sports coaching: Research and practice. Routledge.
  42. US Center for SafeSport (2020). 2020 Athlete culture and climate survey.
  43. Van Woezik, R. A., McLaren, C. D., Côté, J., Erickson, K., Law, B., Horning, D. L., … & Bruner, M. W. (2021). Real versus ideal: Understanding how coaches gain knowledge. International Sport Coaching Journal9(2), 189-202.
  44. Vella, S., Oades, L., & Crowe, T. (2011). The role of the coach in facilitating positive youth development: Moving from theory to practice. Journal of Applied Sport Psychology23(1), 33-48.
  45. Vertommen, T., Kampen, J., Schipper-van Veldhoven, N., Wouters, K., Uzieblo, K., & Van Den Eede, F. (2017). Profiling perpetrators of interpersonal violence against children in sport based on a victim survey. Child Abuse and Neglect, 63, 172–182.
  46. Wiersma, L. D., & Sherman, C. P. (2005). Volunteer youth sport coaches’ perspectives of coaching education/certification and parental codes of conduct. Research quarterly for exercise and sport76(3), 324-338.
  47. Willson, E., Kerr, G., Battaglia, A., & Stirling, A. (2022). Listening to athletes’ voices: national team athletes’ perspectives on advancing Safe Sport in Canada. Frontiers in Sports and Active Living4, 840221.
  48. Willson, E., Kerr, G., Stirling, A., & Buono, S. (2022). Prevalence of Maltreatment Among Canadian National Team Athletes. Journal of Interpersonal Violence37(21–22), 1-23.
  49. Wilson, A. L., & Rhind, D. J. (2022). Tracking progress towards the International safeguards for children in sport. Social Sciences11(8), 322.
2024-06-20T12:01:59-05:00June 21st, 2024|General, Research, Sport Education, Sports Coaching, Sports Exercise Science|Comments Off on Coaches’ Perspectives of the Influence of Safe Sport-Related Education

Celebrating the Olympics

A note from the editor: In recognition of the upcoming Olympics, The Sports Journal has “temporarily” allowed for the addition of unique perspectives on Olympic Sports. Please enjoy the commentary from Dr. John Cairney from the University of Queensland.

For the first time in over a decade, NHL players are set to return to the Olympic stage, sparking widespread excitement among ice hockey enthusiasts worldwide. Announced by NHL Commissioner Gary Bettman , this decision to participate in the 2026 and 2030 Winter Games ends a hiatus that has lasted since 2014. It reflects a strategic move to enhance international competition among the world’s elite hockey players, aiming to alternate between the Olympics and the World Cup of Hockey every two years.

The NHL’s withdrawal after the 2014 Olympics stemmed from logistical and financial concerns, including potential revenue losses and the risks of competitive imbalance and player injuries when resuming the season. The injury of John Tavares during the 2014 Sochi Olympics underscored the risk of injury, while also pointing to the demanding nature of Olympic play. Conversely, the break offered a rest period for those not participating, leading to concerns about unequal player fatigue and readiness. Players not competing in the Olympics could potentially benefit from the break, gaining an edge over those who did participate.

Despite these concerns, there was scant research at the time to evaluate their validity, even though professional sports, including ice hockey, are rich in data capable of informing such analyses. Our research team aimed to fill this gap by investigating the impact of NHL participation in the Winter Games on both team and individual player performance, with a
focus on injury and fatigue. Our findings offered some surprising insights.

Our first study looked at the team-level “fatigue effect,” suggesting that teams with more Olympic participants might experience a dip in performance post-Games due to player fatigue, potentially affecting their regular season play. We analysed goal differentials (goals for minus goals against) before and after the Olympics, taking into account the number of players each team sent and mid-season trades’ impacts. Although some Olympic years showed a trend towards a negative effect on goal differential, indicative of a potential fatigue effect, the overall impact on team performance was minor.

The second study focused on individual player performance, particularly during the 2014 Sochi Winter Olympics. We examined performance metrics before and after the Olympics to test the “fatigue theory” at an individual level. Our findings indicated that the number of Olympic minutes played had no significant effect on post-Olympic performance for players overall. However, a closer look at player positions revealed that forwards experienced a slight decrease in points per game post-Olympics if they played more minutes. Defensemen, on the other hand, were unaffected. Overall, our research suggests that concerns about performance declines due to Olympic participation may have been exaggerated.

Our studies provide reassurance that NHL players’ return to the Winter Olympics is beneficial for the sport. While issues related to scheduling, injury risks, and competitive balance remain, the evidence indicates that these factors minimally impact the league and its athletes. The advantages of Olympic participation, including sport promotion, player experience, and fan engagement, significantly outweigh the potential downsides. As the NHL sends its stars back to the Olympic ice, this move is celebrated not only by fans but also as a victory for the global prestige of ice hockey.



2024-05-08T11:05:32-05:00June 7th, 2024|Commentary, General|Comments Off on Celebrating the Olympics

For the Good of the Game: What Keeps Soccer Referees from Renewing Their Licenses

Authors: Dr. J Ross Pruitt1, Dr. Dexter Davis2

Corresponding Author:

J. Ross Pruitt* Professor Department of Agriculture, Geosciences, and Natural Resources

269 Brehm Hall University of Tennessee at Martin

Martin, Tennessee 38238

Phone: (731)881-7254 Fax: (731)881-7968

For the Good of the Game: What Prevents Soccer Referees from Renewing Their Licenses 


The United States faces a critical shortage of youth sports referees despite a growing interest in many sports. This issue is increasingly gaining attention from sanctioning bodies, referee associations, and researchers. There is a significant cost of referee turnover and implementing strategies implemented to increase retention of officials, especially in soccer. Correct identification of the issues resulting in non-renewal of referee licenses will increase the likelihood of retention strategies being successful. This study builds on existing research by using best-worst scaling to provide a preference share on the factors that result in non-renewal which Likert scales cannot provide. Current and former U.S. Soccer Federation referees in Tennessee were surveyed to determine which factors are most likely to motivate their decision to not renew their referee license. Findings from this research indicate that motivations are different from youth referees compared to adult referees. Youth referees find the cost of refereeing and assigning are resulting in non-renewal of licenses compared to lack of respect and changing work commitments among adult referees. Results of this research can be used to improve retention strategies across age groups.

Keywords: best-worst scaling, soccer referees, referee motivations, referee retention  

Organized sports are an important part of society within the United States as it allows recreational and entertainment opportunities for participants and spectators. Sports officials are often referred to as the “third team” and are a critical aspect to the success of organized sports. In recent years, the popular press has been bombarded with stories of referee shortages (e.g., Conlon, 2022; Medina, 2022; Yurkevich, 2023) and physical attacks (Mendola, 2014; Ortiz, 2015; Weir, 2022; Hamacher, 2023). A majority of states have enacted or are considering laws to protect referees according to the National Association of Sports Officials (NASO) who tracks the status of legislation impacting sports officials (NASO, n.d.).  

Even with increased awareness of the issues of referee shortages, verbal abuse and/or physical assaults, and growing legal protections, organized sports in the United States are still facing a shortage of officials. National and grassroots sport associations have enacted strategies to reduce the turnover to aid in recruitment (Titlebaum et al., 2009) and retention (Warner et al., 2013) of sports officials. These efforts will take time to minimize the impact of verbal abuse and physical assaults that are believed to result in the exodus of sports officials (Warner et al., 2013; Downward et al., 2023). Prior research has explored the factors that result in individuals deciding to become a sports official (Furst, 1991; Kellett and Warner, 2011; Johansen, 2015; Baldwin and Vallance, 2016) and continuing as a sports official (Rainey, 1999; Rainey and Hardy, 1999; Kellett and Shilbury, 2007; Kellett and Warner, 2011; Cuskelly and Hoye, 2013; Ridinger et al., 2017; Da Gama et al, 2018; Giel and Brewer, 2020; Orviz-Martinez et al, 2021; Downward et al., 2023), but the factors resulting in non-renewal of licenses needed to officiate is less clear in the literature.  

The internal and external factors that draw individuals to officiate sports are important motivators to keep renewing their license. When one or more of these factors dissipate or change, an official’s lagging desire to continue can result in non-renewal of the soccer refereeing license. Licenses to officiate soccer are typically renewed annually which requires a conscious decision to continue or not continue. This provides the official with the opportunity to reflect whether the benefits of officiating (e.g., financial, health, social) continue to exceed the costs (e.g. cost to renew the license, additional time away from family, job stress, verbal abuse). As very few soccer referees can rely financially on officiating income alone, the need to balance family, career, and officiating is present. The popularity of youth soccer results in a constant cadre of referees needing recruitment, introductory and advanced training, and retention at the youth and grassroots level. Past research (e.g., Gomes et al, 2021) has used Likert scales and qualitative interviews to determine factors that impact continued refereeing of soccer. This study adds to the existing literature by inviting current and former soccer officials to make a choice among the alternative factors included on the survey instrument. The method used in this study presents a direct ranking of factors not provided in Likert scales. This paper continues with a literature review of the existing literature of factors attracting individuals to officiate sports and what results in the decision to no longer referee followed by a description of our survey methodology. Our survey population included current and former U.S. Soccer Federation referees. Results are then discussed with suggestions for future research presented.  

Literature Review  

The reasons an individual becomes a sports official are complex, but often include altruistic motivations (Balch and Scott, 2007) and love of the sport (Burke, Joyner, Pim, and Czech, 2000). Furst (1991) and Balch and Scott (2007) state that officials continue to officiate for social and interpersonal reasons along with a commitment to the sport. Kellett and Shilbury (2007) discuss the importance of the social and interpersonal support provided between officials to cope with the stress of officiating sports. The stress is, in part, a reflection of the need to quickly and correctly apply the rules of the sport while being in the proper position to make a decision. Initial training of new sports officials often focuses primarily on knowing the rules of the sport with some field training to practically apply what is learned. Factors that are important to keep beginning officials engaged in officiating such forming interpersonal relationships (e.g., Furst, 1991; Balch and Scott, 2007; Kellett and Shilbury, 2007; Kellett and Warner, 2011; Baldwin and Vallance, 2016) and coping with stress (e.g., Voight, 2009) are not the primary focus of initial trainings.  

Officiating sports is a stressful experience due to the complexity of making quick decisions (Guillén and Jiménez, 2001; González-Oya, 2006; Gama et al., 2018) in an environment where positive feedback for correct decisions is limited. In younger and/or inexperienced officials, the lack of experience in these environments and ability to cope with the accompanying stress can contribute to referees no longer officiating (Cuskelly and Hoye, 2013). Prior research has focused on the connection between stressors and burnout (Rainey and Hardy, 1999; Voight, 2009; Da Gama et al., 2018; Gomes et al., 2021; Orviz-Martinez et al., 2021; Downward et al., 2023) with tools like the Burnout Inventory for Referees developed by Weinberg and Richardson (1990). Stressors experienced by sports officials are not always related to the sporting event but can be representative of other factors in their lives including work, family, and support of the organization for which they officiate (Voight, 2009; Cuskelly and Hoye, 2013).  

Reasons that individuals begin refereeing may not always be the reasons they intend to continue. Kellett and Shilbury (2007) document that the interpersonal relationships developed can overcome nervousness experienced by beginning officials. These interpersonal relationships can be a positive stressor, or an indication of commitment described in Cuskelly and Hoye (2013). These may be social in nature can result in officials who, “are likely to feel somewhat compelled to continue officiating through various social mechanisms” (Cuskelly and Hoye, 2013). The level of organizational support, or the official’s perception of support, can result in an intention to continue officiating (Rainey, 1999; Kellett and Warner, 2011).  

Giel and Breuer (2020) find the altruistic motives are not a significant factor in continuing to referee. This highlights the importance of the social relationships as the stress associated with officiating, balancing family, job, and officiating, the stress associated with maintaining the desired level of performance, or other factors can result in the official questioning their desire to continue. This contributes to the belief often expressed in the popular press that burnout and verbal abuse/physical assault are primary motivators to officials leaving the sport (Kellett and Shilbury, 2007). The ability to reframe the abuse as described in Kellett and Shilbury (2007) may limit the extent to which the perception is reality. Voight (2009) finds the conflict between family and officiating, making a controversial call, conflict between work and officiating, making the wrong call, and verbal abuse from coaches as the top stressors among college soccer officials. The least amount of stress can be attributed to the fear of physical harm (Voight, 2009).  


The decision to not renew one’s soccer referee license reflects the costs of continuing to referee (whether financial, social, or emotional) relative to the benefits accrued by refereeing. We hypothesize that referees will consider not renewing their license prior to the actual decision where the license is not renewed (Rainey and Hardy, 1999; Cuskelly and Hoye, 2013). Factors that motivate the decision to not renew one’s license are presented in Table 1. Included factors represent those included in the literature (e.g., Furst, 1991; Rainey, 1999; Rainey and Hardy, 1999; Burke et al., 2000; Balch and Scott, 2007; Kellett and Shilbury, 2007; Cuskelly and Hoye, 2013; Johansen, 2014; Giel and Breuer, 2020) as well as those from our personal experiences refereeing and coaching soccer. After the factors shown in Table 1 were selected to include in the questionnaire, the staff and mentors of the U.S. Youth Soccer Region III Championships reviewed our factors and accompanying descriptions for thoroughness. Their suggestions are reflected in our final factors presented in Table 1.  

Use of best-worst scaling (Finn and Louviere, 1992) provides the relative importance that a factor can have on a referee’s continued interest in renewing their license. This method provides an improvement over qualitative interviews which can provide insight into motivations for referees, but not a hierarchical preference ranking that can be used by referee associations to assist in retention of referees. An additional benefit of best-worst scaling is the fact it provides a ratio scale for its results unlike a Likert rating scale that may result in the ordinal ranking not being consistent across respondents (Steenkamp and Baumgartner, 1998; Lusk and Briggeman, 2009). This provides greater insight into the obstacles for a referee to annually renew their license which can lead to increased retention efforts and educational efforts by clubs and sanctioning bodies to reduce the impact of factors resulting in non-renewal of licenses. 

Best-worst scaling provides the respondent the ability to select the factor that provides the most and least utility in a choice set which Likert scales do not provide. This approach has significant implications for marketing (Cohen, 2009; O’Reilly and Huybers, 2015; Massey, Wang, and Waller, 2015) to help identify specific factors that consumers find desirable. Use of this method has extended into the healthcare industry (Flynn et al., 2007) and the value of public information (Pruitt et al., 2014). Given J factors, there are J(J-1) combinations a respondent could select for each best-worst question. The choice of the most important factor j by individual i can by represented by λj on the utility scale with the latent level of utility determined by Iij = λj + εij which assumes that εij is the random error term. By selecting factor j as the most important factor and factor k as the least important is determined by the probability for all other J(J-1)-1 possible differences in the choice set.  

Results from best-worst scaling normally occurs through a multinomial or random parameters logit. Estimate coefficients have little interpretation aside from the magnitude of the coefficient. Preference shares for each factor’s impact on lack of interest in continuing to referee is calculated using the following equation preference share for factor.

Respondents were asked if they had actively considered not renewing their U.S. Soccer Federation (USSF) referee license in the past five years. Individuals that responded yes, were then asked best-worst questions using the factors that were identified and presented in Table 1. Using PROC OPTEX in SAS 9.4, a quasi-balanced incomplete block design (BIBD) was created. The design had a treatment D-efficiency of 90.78 and a block design D-efficiency of 99.86. This resulted in twelve best-worst questions with six factors present in each question. Each factor appeared six times to each respondent with an example of the best-worst questions is provided in Figure 1.

Figure 1. Example Best-Worst Question


A web-based Qualtrics survey was created that was distributed to current and former U.S. Soccer Federation referees implementing the best-worst questions discussed previously. Through contacts with the Tennessee Soccer Referee Program, we were able to distribute the questionnaire to 3,507 current and former referees. Our ability to contact referees who had not recertified in the previous four years is due to the Tennessee Soccer Referee Program adopting computer software that allows the program to track referees who do not re-certify from year to year. Inclusion of youth referees (between the ages of thirteen and eighteen) was approved by our university’s Institutional Review Board which allows for determination if factors vary by age. Per USSF policy, any email contact from a certified USSF assignor results in the parent/guardian also being contacted1. This resulted in parents/guardians of current and former youth referees also receiving the recruitment email. Initial questions identified if the respondent was at least eighteen years of age and then determined if the respondent was answering for themselves or as parent/guardian of a current or former youth soccer referee2. For youth referees, we included questions that determined if their parent/guardian had provided consent in addition to the minor providing assent. As the parent/guardian also received the recruitment email, email addresses for minors were collected in case the parent/guardian revoked consent necessitating removal of youth referee responses. No parent or guardian contacted us requesting removal of the youth referee’s responses.

A recruitment email was sent in early March 2023 to 3,507 current and former referees registered with USSF in the state of Tennessee with a follow-up email sent two weeks later. An incentive was offered to each respondent of a gift card worth $100 to a referee equipment supplier or a free registration for the 2023 year. Email addresses were collected at the end of the questionnaire and provided to the Tennessee Soccer Referee Program which was responsible in selecting and contacting the winners of the inducement. We received 107 usable responses for a response rate of 3.05%.

Results Demographic information is provided in Table 2. Total responses did vary by question as respondents were not required to answer every demographic question which were asked following the best-worst questions. Respondents were overwhelmingly male and Caucasian. Approximately forty percent of respondents were less than twenty-five years of age and an additional twenty-five percent between the ages of forty-three and fifty-four. Over sixty percent who responded were no longer refereeing soccer with approximately two-thirds believing they were assigned the appropriate number of matches given their skill and ability level. Those receiving the questionnaire were asked an open-ended question on how many years they refereed soccer. Of the 110 responses, many did not provide an exact number. For those who provided an exact number, the average number of years that survey participants had refereed was 8.63 years. Given responses not included in this calculation that stated they had refereed 10+, 20+, or 50+ years, this estimate of 8.63 understates the longevity of referees in this research. A histogram of responses for this question is presented in Figure 2. More than three-quarters of respondents refereed no more than sixty matches a year with the majority refereeing less than fifteen matches annually. Over ninety percent of respondents only refereed soccer. Nearly seventy percent of respondents had suffered verbal abuse in the past two years with approximately five percent having suffered a physical assault (e.g., touched, pushed, shoved, punched, kicked, or spat on) by a player, coach, fan, or parent. Parents and coaches were most likely to have been the source of verbal abuse with players being the source of physical assault.

As we were able to include youth referees (less than eighteen years old), we conducted t-tests for significant differences in means between those who had actively considered not renewing their USSF licenses for youth and adult referees. We did not test for differences in means in age and educational attainment categories since we compared those less than eighteen of ages to all other ages in this comparison. Differences in the mean at the 5% level of significance (p<0.05) were found in these groupings with less than fifteen matches officiated, whether the respondent felt they were under assigned, assigned the right number of matches for their skill/ability level, and whether they play organized soccer. Table 2 includes these results.

Table 2. Demographic Information

VariableMeanStandard Deviation
Gender (n=111)  
Prefer Not to Say0.90%0.09
Ethnicity (n=111)  
African American0.00%0.00
Native American0.00%0.00
Prefer Not to Say5.41%0.23
Age (n=111)  
Over 6010.81%0.31
Education Level (n=111)  
Currently in Middle/High School27.03%0.45
High School Diploma or GED0.00%0.00
Trade, vocational, or technical school4.50%0.21
Associate Degree4.50%0.21
Bachelor’s Degree27.93%0.45
Master’s Degree15.32%0.36
Doctoral or Professional Degree7.21%0.26
Prefer Not to Say1.80%0.13
Household Income (n=110)  
Less than $40,00010.00%0.30
$40,000 to $60,0009.09%0.29
$60,001 to $80,0008.18%0.28
$80,001 to $100,0005.45%0.23
Greater than $100,00040.00%0.49
Prefer Not to Say27.27%0.45

Table 2. Continued

VariableMeanStandard Deviation
Residence (n=111)  
Urban Area14.41%0.35
Suburban Area66.67%0.39
Rural Area18.92%0.47
Levels Officiated1  
Youth recreational33.46% 
High School16.18% 
Adult Amateur/Recreational10.29% 
Approximate number of annual matches  
Less than 1530.91%20.46
Over 1057.27%0.26
Proper Assigning Level (n=109)  
Under assigned25.69%20.44
Over assigned7.34%0.26
Right number66.97%20.47
Sports Officiated besides Soccer  
5 or more0.00%0.00
Play Organized Soccer (n=110)43.64%20.50
Verbally Abused in Last Two Years (n=109)68.81%0.47
Source of Verbal Abuse1  

Table 2. Continued

VariableMeanStandard Deviation
Physically Assaulted in Last Two Years (n=109)4.59%0.21
Source of Physical Assault1  
Injury of at Least Four Weeks (n=109)11.93%0.33
Attend Continuing Education (n=110)  
Once a year28.18%0.45
Twice a year7.27%0.26
Three to four times a year10.00%0.30
At least five times a year0.00%0.00
Does not attend47.27%0.50
Accepts unsanctioned matches (n=110)12.73%0.33
Anticipates refereeing soccer: (n=110)  
No longer refereeing60.91%0.49
Less than one year7.27%0.26
One to two years12.73%0.33
Three to four years8.18%0.28
At least five years10.91%0.31

1 Question allowed multiple responses and standard deviations are not presented as a result.
2 Denotes significant differences at the 5% level (p<0.05) between youth and adult referees who had actively considered not renewing their license.

Non-Renewal of Referee License

Respondents who answered they had actively considered not renewing their license in the past five years were shown a series of questions asking them to select the most and least important factors impacting why they would not renew their refereeing license. As our sample included youth referees (those less than 18 years of age), we estimated a combined model for all referees responding against the alternative models of youth and adult referees. Each of these models was estimated using a multinomial logit (MNL), an uncorrelated random parameters logit (RPL), and a correlated random parameters logit model. Significant differences were found to exist between youth and adult referees who were considering not renewing resulting in separate models being estimates for youth and adult referees. Likelihood ratio tests favored the use of MNL model for both youth and adult referees.

Youth Referees

Results for youth referees are presented in Table 3 uses Work as the base factor with results. Estimates for the MNL and RPL models are presented with the MNL preferred by use of a likelihood ratio test. Aside from their magnitude, the econometric estimates in Table 3 have no natural interpretation and equation 1 was used to calculate the shares of preference that are presented. The shares of preference for the uncorrelated RPL model were generated from 1,000 random draws using a normal distribution of the mean and standard deviation of a specific factor that might result in a referee not renewing their USSF license. Shares of preference were consistent between the two modeling techniques as there was not greater than ±0.01% difference for any factor. The cost to referee (i.e., Afford) was the number one reason that youth referees had considered not renewing their USSF license. This factor includes the inability to make it to matches for youth referees reflecting the need for an adult or friend to help them make it to assignments. Note that even with a small sample size of youth referees, fifteen of the eighteen youth referees were no longer refereeing. The youth referee’s opinion on how well they were assigned was the second most important factor with the lack of Respect from fans, players, and coaches third (depending on the model used). It should be noted that the fourth most important factor was Game Fees, indicating the cost to benefit ratio for youth referees is contributing to non-renewals. The use of best-worst scaling provides a clearer view of the magnitude of factors resulting in youth referees not renewing their licenses through the direct comparisons with the lack of Respect relatively not as important as other factors.

Table 3.  Relative Importance of Factor Impacting Non-Renewal of Youth Referee Licenses

FactorEconometric EstimatesShares of Preference
Youth Involvement-0.761***-0.763***0.0270.026
Social Aspects0.271***0.2700.0740.074
Family Commitments-0.865***-0.859***0.0240.024
Lack of Opportunities to Advance-0.189-0.1880.0470.047
Cost to Referee1.123****1.130***0.1740.174
Game Fees0.913***0.915***0.1410.141
Lack of Organizational Support0.524**0.529**0.0960.096
(Base Factor)  [0.000][0.015]
Log Likelihood-625.138-624.910  
McFadden’s LRI0.0940.149  
Number of Respondents1818  

            ***, **, and * asterisks represent the factor is significantly different from the Work factor at the 1%, 5%, and 10% level, respectively.

a Numbers in parentheses are standard errors.
b numbers in brackets are standard deviations.

In addition to the shares of preference presented in Table 3, we generated Pearson correlations from the individual specific RPL estimates shown in Table 4. Several factors had correlations with at least ±0.3 with another factor. Given the limited number of responses, care should be taken when viewing Table 4, but it provides an indication of how youth referees view these factors influencing their decision to not continue refereeing. The more likely a youth referee viewed the lack of Social camaraderie, the higher an injury might factor into a non-renewal decision. Importantly, the lack of Social connections had a strong direct relationship with their views of Organizational Support provided to them. Concerns about how many games the referee was assigned had a positive relationship with Game Fees being an important factor in the decision to not renew the license. Game Fees tended to have large (positive or negative) correlations with many factors that were included in this research.

Table 4.  Pearson Correlations Between Factors from Individual Specific RPL Estimates of Youth Referees

Respect (1)1.000          
Youth Involvement (2)0.0071.000         
Assign (3)-0.1400.1771.000        
Social (4)-0.1760.410-0.3011.000       
Injury (5)-0.007-0.439-0.4940.5051.000      
Advance (6)-0.211-0.8320.0180.2220.2121.000     
Age (7)0.1640.8740.100-0.248-0.448-0.9191.000    
Cost (8)-0.043-0.159-0.2640.2270.079-0.0730.0961.000   
Game Fees (9)-0.1230.7250.476-0.569-0.737-0.5380.694-0.0461.000  
Organizational Support (10)0.097-0.338-0.1720.5140.5440.189-0.3320.249-0.6691.000 
Family (11)-0.326-0.753-0.0300.4890.5640.723-0.860-0.025-0.6100.5311.000

Adult Referees

Results for adult referees who had considered not renewing their USSF license are presented in Table 5. As with youth referees, a MNL model was preferred to an uncorrelated RPL model with the estimates from both models presented. Unlike youth referees, the lack of Respect experienced by adult referees is the primary reasons resulting in the non-renewal decision. Work commitments or a change in them was the second most important factor. Nearly two-thirds of adult referees who had considered not renewing their license were no longer refereeing; fifteen were considering not renewing in more than the next two years with only four considering refereeing at least four more years.

Table 5.  Relative Importance of Factor Impacting Non-Renewal of Adult Referee Licenses

FactorEconometric EstimatesShares of Preference
Youth Involvement-1.571***-1.582***0.0280.028
Social Aspects-1.313***-1.312***0.0360.036
Family Commitments-0.688***-0.688***0.0680.068
Lack of Opportunities to Advance-0.828***-.833***0.0590.059
Cost to Referee-.481***-0.474***0.0840.084
Game Fees-0.350***-0.348***0.0950.095
Lack of Organizational Support-0.510***-0.514***0.0810.081
(Base Factor)  [0.000][0.014]
Log Likelihood-2843.485-2838.531  
McFadden’s LRI0.0660.097  
Number of Respondents7777  

***, **, and * asterisks represent the factor is significantly different from the Work factor at the 1%, 5%, and 10% level, respectively.

a Numbers in parentheses are standard errors.
b numbers in brackets are standard deviations.

As with the youth referees, Pearson correlations for the adult referees are presented in Table 6. A greater response rate among adults compared to youth referees provides more robustness in the correlations that are presented. It is interesting to note the strong negative correlation between Game Fees and Assign (-0.591) suggesting concerns about pay is not tied to assigning. Concerns about Game Fees and the ability to Advance had a strong positive relationship (0.607) indicating adult referees view the pay for higher level games isn’t a strong enough incentive to advance. Those referees who rated the inability to Advance highly was negatively correlated (-0.611) with concerns about being over or under assigned (Assign).

Table 6.  Pearson Correlations Between Factors from Individual Specific RPL Estimates of Adult Referees

Respect (1)1.000          
Youth Involvement (2)0.0631.000         
Assign (3)-0.304-0.5781.000        
Social (4)-0.105-0.0520.0171.000       
Injury (5)0.1530.533-0.587-0.1241.000      
Advance (6)0.3000.387-0.6110.2170.4181.000     
Age (7)-0.172-0.2300.089-0.192-0.115-0.2871.000    
Cost (8)0.2750.179-0.4600.0280.1960.542-0.2821.000   
Game Fees (9)0.0460.444-0.591-0.1240.4010.607-0.0150.3361.000  
Organizational Support (10)0.129-0.036-0.055-0.0690.049-0.334-0.0930.007-0.1971.000 
Family (11)-0.255-0.2790.574-0.136-0.543-0.3380.147-0.483-0.3230.0361.000


Concerns about retaining sports officials are a pressing factor for many sports with referee abuse a concern among leagues and official associations. Factors influencing the decision to not renew referee licenses are not well understood in the literature. Prior research has focused on qualitative factors impacting the renewal decision which doesn’t quantitatively rank factors included in the research. This research surveyed current and former referees who had actively considered not renewing their referee license with a majority no longer refereeing soccer. There were significant differences between youth and adult referees in the factors that had led them to consider not renewing their referee license. For youth, the cost to referee and concerns about being over- or under-assigned were the top two reasons for considering not renewing their license compared to adults who were more concerned about the lack of respect and work commitments. For both age groups, concerns about organizational support were significant factors as it relates to continuing refereeing.

Our study is limited by the small sample size, but it is an important look into the factors that resulted in a majority of referees no longer renewing their U.S. Soccer Federation license. While we do not focus on the well-being of referees as in Downward and Webb (2023), our findings are consistent with theirs that a zero-tolerance approach will aid in adult referee retention. This reinforces the need for organizational support (Rainey, 1995; Voight, 2007; Ridinger et al., 2017; Downward and Webb, 2023), but also requires training by those organizations on what to include in post-match reports to have the backing. As over 75% of respondents in our survey did not attend more than one continuing education session annually, sanctioning bodies and referee associations need innovative ideas to aid in reaching this objective.

Future research should focus on expanding this to referees who have not recently considered non-renewal of their referee licenses. This portion of the referee community will likely have different factors motivating their continued renewals as was demonstrated by the differences observed in this paper based on the age of the referee. Identification of the factors that aid in retention of these referees may aid in development of strategies to limit the impact of factors discussed in this research. Given the nature of soccer in the U.S., future research should better control for the differences in length of refereeing and level officiated (e.g., recreational versus club). With the number of young referees who work matches in the U.S., the skills necessary to be successful may not have been developed to handle the stressors commonly associated with officiating (Rainey, 1995; Rainey and Hardy, 1999; Burke et al., 2000; Voight, 2009; Gomes et al., 2021). A more diverse respondent pool, in terms of locality, gender, and ethnicity, is also needed to better understand why referees continue to engage in a stressful avocation.


The authors express appreciation to Don Eubank, State Referee Administrator for Tennessee Soccer, for sending the questionnaire to soccer referees in the state and providing the incentive for respondents to complete the questionnaire. We also thank the staff and mentors of U.S. Youth Soccer Region III for helpful feedback on an early draft of the questionnaire. The authors are grateful for the helpful edits and suggestions from Marco Palma on an earlier draft of this paper.

Conflicts of Interest

J. Ross Pruitt is an active soccer referee with the U.S. Soccer Federation, Tennessee Secondary School Athletic Association, and National Intercollegiate Soccer Official Association.


  1. Balch, M. J., & Scott, D. (2007). Contrary to Popular Belief, Refs are People Too! Personality and Perceptions of Officials. Journal of Sport Behavior, 30(1).Baldwin, Christopher, and Roger Vallance. “Rugby Union Referees’ Experiences with Recruitment and Retention,” n.d.
  2. Burke, K. L., Joyner, A. B., Pim, A., & Czech, D. R. (2000). An exploratory investigation of the perceptions of anxiety among basketball officials before, during, and after the contest. Journal of Sport Behavior, 23(1).Cohen, E. (2009). Applying best-worst scaling to wine marketing. International Journal of Wine Business Research, 21(1), 8-23.
  3. Conlon, C. (2022, September 13). Years-long Montana referee shortage getting worse as cancellations loom. Q2 News (KTVQ). Accessed June 13, 2023.
  4. Cuskelly, G., & Hoye, R. (2013). Sports officials’ intention to continue. Sport Management Review, 16(4), 451-464.
  5. Da Gama, D. R. N., Nunes, R. D. A. M., Guimarães, G. L., Leandro De Lima, E. S., De Castro, J. B. P., & Vale, R. G. D. S. (2018). Analysis of the burnout levels of soccer referees working at amateur and professional leagues of Rio de Janeiro, Brazil. Journal of Physical Education and Sport, 18, 1168-1174.
  6. Downward, P., Webb, T., & Dawson, P. (2023). Referee abuse, intention to quit, and well-being. Research quarterly for exercise and sport, 1-11.
  7. Finn, A., & Louviere, J. J. (1992). Determining the appropriate response to evidence of public concern: the case of food safety. Journal of Public Policy & Marketing, 11(2), 12-25.
  8. Flynn, T. N., Louviere, J. J., Peters, T. J., & Coast, J. (2007). Best–worst scaling: what it can do for health care research and how to do it. Journal of health economics, 26(1), 171-189.
  9. Furst, D. M. (1991). Career contingencies: Patterns of initial entry and continuity in collegiate sports officiating. Journal of Sport Behavior, 14(2), 93.
  10. Giel, T., & Breuer, C. (2020). The determinants of the intention to continue voluntary football refereeing. Sport Management Review, 23(2), 242-255.
  11. Gomes, A. R., Fontes, L. M. C., Rodrigues, M., & Dias, B. (2021). Burnout in referees: Relations with stress, cognitive appraisal, and emotions. González-Oya, J. Psicología Aplicada al Árbitro de Fútbol: Características Psicológicas y su Entrenamiento; Wanceulen: Sevilla, Spain, 2006.
  12. Guillén, F., & Feltz, D. L. (2011). A conceptual model of referee efficacy. Frontiers in psychology, 2, 25.
  13. Guillén García, F., & Jiménez Betancort, H. (2001). Características deseables en el arbitraje y el juicio deportivo. Revista de psicología del Deporte.
  14. Hamacher, B. (2023, January 25). Soccer player arrested after video shows him attacking referee in Kendall: Police. NBC 6 South Florida. Accessed June 13, 2023.
  15. Jacobs, B. L., Tingle, J. K., Oja, B. D., & Smith, M. A. (2020). Exploring referee abuse through the lens of the collegiate rugby coach. Sport Management Review, 23(1), 39-51.
  16. Johansen, B. T. (2015). Reasons for officiating soccer: the role of passion-based motivations among Norwegian elite and non-elite referees. Movement & Sport Sciences-Science & Motricité, (87), 23-30.
  17. Kellett, P., & Shilbury, D. (2007). Umpire participation: Is abuse really the issue?. Sport Management Review, 10(3), 209-229.
  18. Kellett, P., & Warner, S. (2011). Creating communities that lead to retention: The social worlds and communities of umpires. European Sport Management Quarterly, 11(5), 471-494.
  19. Lusk, J. L., & Briggeman, B. C. (2009). Food values. American journal of agricultural economics, 91(1), 184-196.
  20. Massey, G. R., Wang, P. Z., Waller, D. S., & Lanasier, E. V. (2015). Best–worst scaling: A new method for advertisement evaluation. Journal of Marketing Communications, 21(6), 425-449.
  21. Medina, E. (2022, April 21). Bad behavior drove a referee shortage. Covid made it worse. The New York Times. Accessed June 13, 2023.
  22. Mendola, N. (2014, July 31). Testimony in death of Michigan referee provides disturbing glimpse into incident – NBC Sports. NBC Sports. Accessed June 13, 2023.
  23. Legislation status – National Association of Sports Officials. (2023, October 12). National Association of Sports Officials. Accessed November 1, 2023.
  24. O’Reilly, N., & Huybers, T. (2015). Servicing in sponsorship: A best-worst scaling empirical analysis. Journal of sport management, 29(2), 155-169.
  25. Referee drilled by two Texas high school football players was a Fill-In. (2015, September 9). NBC News. Accessed June 13, 2023.
  26. Orviz-Martínez, N., Botey-Fullat, M., & Arce-García, S. (2021). Analysis of burnout and psychosocial factors in grassroot football referees. International Journal of Environmental Research and Public Health, 18(3), 1111.
  27. Pruitt, J. R., Tonsor, G. T., Brooks, K. R., & Johnson, R. J. (2014). End user preferences for USDA market information. Food Policy, 47, 24-33.
  28. Rainey, D. (1995). Sources of stress among baseball and softball umpires. Journal of Applied Sport Psychology, 7(1), 1-10.
  29. Rainey, D. W. (1999). Sources of stress, burnout, and intention to terminate among basketball referees. Journal of sport behavior, 22(4), 578-590.
  30. Rainey, D. W., & Cherilla, K. (1993). Conflict with baseball umpires: An observational study. Journal of Sport Behavior, 16(1), 49-60.
  31. Rainey, D. W., & Hardy, L. (1999). Sources of stress, burnout and intention to terminate among rugby union referees. Journal of Sports Sciences, 17(10), 797-806.
  32. Ridinger, L. L., Warner, S., Tingle, J. K., & Kim, K. R. (2017). Why referees stay in the game. Global Sport Business Journal, 5(3), 22.
  33. Steenkamp, J. B. E., & Baumgartner, H. (1998). Assessing measurement invariance in cross-national consumer research. Journal of consumer research, 25(1), 78-90.
  34. Titlebaum, P. J., Haberlin, N., & Titlebaum, G. (2009). Recruitment and retention of sports officials. Recreational Sports Journal, 33(2), 102-108.
  35. Voight, M. (2009). Sources of stress and coping strategies of US soccer officials. Stress and Health: Journal of the International Society for the Investigation of Stress, 25(1), 91-101.
  36. Warner, S., Tingle, J. K., & Kellett, P. (2013). Officiating attrition: The experiences of former referees via a sport development lens. Journal of Sport Management, 27(4), 316-328.
  37. Weinberg, R. S., & Richardson, P. A. (1990). Psychology of officiating. Leisure Press.
  38. Weir, G. (2022, September 10). Arkansas High School Football Game Put On Pause After Elderly Fan Tries To Attack Referee. OutKick. Accessed June 13, 2023.
  39. Wolfson, S., & Neave, N. (2007). Coping under pressure: Cognitive strategies for maintaining confidence among soccer referees. Journal of Sport Behavior, 30(2), 232-247.
  40. Yurkevich, V. (2023, May 18). America has an umpire shortage. unruly parents aren’t helping | CNN Business. CNN. Accessed June 13, 202
2024-05-08T12:27:56-05:00May 31st, 2024|General, Sports Coaching, Sports Management|Comments Off on For the Good of the Game: What Keeps Soccer Referees from Renewing Their Licenses

Advice on making the most of basketball three-point shot data

Authors: George Terhanian1

Corresponding Author:

George Terhanian, PhD
200 Hoover Avenue, #2101
Las Vegas NV, 89101

1George Terhanian founded Electric Insights after holding executive positions at The NPD Group, Toluna, and Harris Interactive. He has also served on boards or advisory groups for several organizations, including the US National Academy of Sciences, the Advertising Research Foundation, and the British Polling Society. He is known for conceiving how to make survey data, including pre-election forecasts, more accurate through statistical matching methods.

Making the most of basketball three-point shot data


This study’s primary goal is to help National Basketball Association (NBA) and other basketball teams worldwide increase their three-point shooting accuracy and decrease their opponents’, a key to winning more games.  A related goal is to explain how a combination of good data, logistic regression analysis, likely effects reporting in probabilities or percentage points, and self-serve simulation can improve communication among data analysts, basketball coaches, and players, and enhance each group’s effectiveness.  Logistic regression analysis of 32,511 NBA three-point shots shows six factors affect the three-point shooting percentage: closest defender’s distance to the shooter, time left on the 24-second shot clock, whether the player shot after dribbling or catching the ball, game period, shot distance, and venue.  In the past, data analysts conveyed the results of such analyses to coaches and players using terms such as regression, logits, and odds.  Some NBA executives say doing so again would be disastrous.  An alternative is to emphasize probabilities and percentages in communication and create self-serve simulators coaches and players can use to predict how changes in critical factors affect three-point shooting percentages.  NBA and other teams worldwide can apply this approach to new and existing datasets they maintain, enhance, and build.

Key Words: self-serve simulation, predicted probabilities, logistic regression, likely effects reporting, psychotherapy


The National Basketball Association (NBA) releases specific three-point shot characteristics, such as shooter name and shot distance.  Aside from the 2014-15 season’s first 903 of 1,230 games (and 2015-16’s first 631, though the latter data are no longer publicly available), the released data exclude a variety of individual shot characteristics such as the closest defender’s distance to the shooter, a crucial defensive effectiveness measure (14).  Teams are said to consider the excluded characteristics proprietary.  As Mike Zarren, assistant general manager and chief legal counsel for the NBA’s Boston Celtics, explained, “You can’t share stuff with other teams…We are not at an equilibrium point where all the teams know what everyone else is doing.  There are some advantages that some teams have over others” (15) (51:47). 

The analyses here use the 2014-15 shot dataset, the last and largest single-season one containing full shot data that is publicly available.  The main goal is to help NBA and other basketball teams worldwide increase their three-point shooting accuracy and decrease their opponents’.  Teams that do so should win more games.  A related goal is to explain how a combination of good data, logistic regression analysis, likely effects reporting in probabilities or percentage points (e.g., “Shooting off the catch rather than the dribble is associated with a two-percentage-point increase in our three-point shot make percentage.”), and self-serve simulation can improve communication among data analysts, basketball coaches, and players, and enhance each group’s effectiveness.  NBA and other teams worldwide can apply this approach to new and existing datasets they maintain, enhance, and build.  Aspects of the approach are also portable to many other issues and areas where the key outcome variable is binary (26).

This paper has seven additional sections (excluding references and other ancillary information).  The first summarizes basic rules and strategies for NBA basketball, highlighting the importance of the three-point shot.  It also explains why data analysts seeking to communicate effectively with coaches and players should consider using non-technical language.  The second section describes the three-point shot data used in this paper’s analyses.  It then provides the rationale for relying on logistic regression analysis for model building and prediction.  The third section reports the results of the analyses and suggests how data analysts might share them with coaches and players.  It also explores why academic researchers tend not to report likely effects in probabilities or percentage points.  The fourth details how data analysts can build self-serve simulators that report likely effects in probabilities or percentage points.  The limitations of this paper’s analyses are discussed in the fifth section.  The next-to-last section describes how teams might apply the approach described here, while the final section provides concluding remarks.

NBA Basketball: Basic Rules and Strategies

NBA games have two teams with five players competing for four 12-minute periods (excluding possible five-minute overtime periods).  To score, a team needs to shoot the ball through the basket.  With the clock running, a successful shot is worth three or two points, depending on the shooter’s distance from the basket.  The clock stops for free throws, which are uncontested 15-foot shots worth a single point awarded for specific infringements.  One can calculate each shot’s expected value (EV) by multiplying its potential value by its average make percentage.  For the 2022-23 regular season, the expected value of a three- and two-point shot was almost identical: 1.08 points (3*.36) for a three-pointer and 1.10 (2*.55) for a two-pointer.  Each free throw’s expected value was .78 points (1*.78) or 1.56 for a more typical pair (3).  A recent example shows why the expected value measure can be strategically important.

In the second round of the 2020-21 playoffs, the Atlanta Hawks shocked the heavily favored Philadelphia 76ers, coming from behind to win the seven-game series four to three.  The Hawks’ decision to foul Ben Simmons repeatedly to force him to shoot free throws contributed to the victory.  As Hall-of-Fame player Earvin “Magic” Johnson observed, “…it fueled the Hawks’ comeback” (13).

Simmons shot just 33% (15 for 45) from the free-throw line for the series, far below his 61% (and the league’s 77%) regular season average.  Simmons’s 33% figure suggests the Hawks expected him to score only .66 points for two free throws in a series in which his team made 40% of its three-pointers (for an expected value of 1.22 points) and 52% of its two-pointers (for a 1.05 expected value).  That means the Hawks expected to gain .56 points (1.22 – .66) for a replaced three-point shot and .39 points (1.05 – .66) for a replaced two-pointer with the foul Simmons strategy.  Perhaps more notably, it may have affected Simmons’s decision-making.  To his team’s detriment, Simmons chose not to attempt an open lay-up or dunk with 3:30 remaining in game seven (4), arguably for fear of getting fouled and having to shoot free throws (21, 27).

Overstating three-point shooting’s significance is difficult.  In 2022-23, the Toronto Raptors, Charlotte Hornets, and Houston Rockets won 41, 27, and 21 (of 82) regular season games, too few to qualify for the post-season playoffs; their three-point shooting percentages of 34%, 33%, and 33% were the league’s worst.  The Philadelphia 76ers, Golden State Warriors, and Los Angeles Clippers won 54, 44, and 44 games, enough to compete in the playoffs; they were top performers in three-point shooting at 39%, 39%, and 38%.  These data and separate multi-season analyses (18, 20) suggest that winning in the NBA hinges heavily on making (and defending) three-point shots. 

Clear Communication 

An excellent statistical model is “a simplified version of reality, like a street map that shows you how to travel from one part of a city to another” (28) (p. ix).  But that map will not help you find your way if it includes esoteric terms or unfamiliar signs or symbols.  Likewise, if data analysts use uncommon language when giving advice, coaches and players may feel lost.  Mike Zarren would agree.  If Celtics’ data analysts were to apply logistic regression to three-point shot data, he would tell them to communicate what they learn “without using the word regression because that’s a disaster” (15) (11:18).  Terms like logits, standard deviations, odds, odds ratios, and z scores also would be off-limits.  Zarren does not believe coaches and players are unintelligent.  Even good data analysts can find aspects of logistic regression challenging.  That is why DeMaris (7) (p. 1,057) observed, “…there is still considerable confusion about the interpretation of logistic regression results.”  And why Gelman and Hill (11) (p. 83) commented, “…the concept of odds can be difficult to understand, and odds ratios are even more obscure.”

Washington Wizards’ assistant coach Dean Oliver’s views on clear communication resemble Zarren’s.  “When I directed quantitative analysis for the Denver Nuggets and would prepare stuff for coaches,” he said, “there were actually very few numbers in there.  It was usually words because it was easier for them to absorb…” (15) (48:54). 

An alternative to avoiding numbers is to report key predictor variables’ likely effects with familiar ones like probabilities and percentages—the NBA reports various descriptive statistics and cross-tabulations on its website, emphasizing percentages, hence coaches’ and players’ familiarity. 



The NBA has used technology to gather detailed player performance data since the 2013-14 season via SportVU, then Second Spectrum.  The analyses here use SportVU data, described as “real-time and innovative statistics based on speed, distance, player separation, and ball possession for comprehensive analysis of players and teams” (25).  How did the SportVU system work?  In each arena’s rafters, six cameras recorded information throughout each game in .04-second intervals, producing 25 images per second.  A computer algorithm then plotted the locations of the ball, basket, and 10 players.  SportVU delivered data and reports to each team and the league as a last step.

As noted earlier, the NBA made available SportVU raw, shot-level data—including the defender distance variable—for three-quarters of the 2014-15 regular season.  (The NBA also made available raw, shot-level data early in the 2015-16 season before discontinuing the practice entirely in January 2016.  The latter dataset is no longer publicly available.)  The 2014-15 dataset (17)—the last and largest single-season one publicly available—contains 21 variables and 128,069 three- and two-point shots, as described in the Appendix.  After making minor changes (e.g., removing two-point shots), the remaining three-point shots totaled 32,511—11,426 makes and 21,085 misses—taken from October 28, 2014, through March 4, 2015.

Analysis Method 

Logistic regression models the relationship between a binary outcome (e.g., made or missed three-point shots, or nearly anything with a yes or no interpretation) and, typically, several predictor or explanatory variables.  It is ideal for identifying and estimating the effects of actions to increase or decrease the size or proportion of the group of interest, specifically, made three-point shots.  It can also predict each three-point shot’s probability of belonging to the “made” rather than the “missed” group.  Many academic researchers consider it “the standard way to model binary outcomes” (11) (p. 79), “dominating all other methods in both the social and biomedical sciences” (2) (para. 1).

The final logistic regression model comprises one dependent and six predictor variables.  The predictor variables were selected based on their relationship with the dependent variable, one another, theory, availability, and their effect on the model’s predictive accuracy.  Below are descriptions of the seven variables and brief explanations for how they may differ from the original ones described in the Appendix.

  1. ShotResult: The dependent variable: whether the shooter made the shot. (Values: 0=Missed, 1=Made; Original variable: Fgm)
  2. DefDist: The closest defender’s distance to the shooter in feet (ft.). Basketball players and coaches recommended a four-category variable after discussions and preliminary analyses. (Values: 1=0-3 ft., 2=3-6 ft., 3=6-9 ft., 4=9+ ft.; Original variable: Close_Def_Dist)
  3. ShotClock: The number of seconds (secs.) on the 24-second shot clock. Analyses showed steep drops in the make probability at the 4- and 2-second marks, thus the decision to create a variable with three categories. (Values: 1=0-2 secs., 2=2-4 secs., 3=4+ secs; Original variable: Shot_Clock)
  4. Catch: Whether the shooter took the shot off the catch or dribble. The original variable reported the number of dribbles the shooter took before shooting. Basketball players and coaches recommended a two-category variable after discussions and preliminary analyses. (Values: 1=Off Catch, 2=Off Dribble; Original variable: Dribbles)
  5. Period: The game period when the shot was taken, with fourth period and overtime shots pooled because of their similar make percentages. (Values: 1=1, 2=2, 3=3, 4=4+; Original variable: Period)
  6. ShotDist: The distance in feet from the center of the basket to the shooter. Basketball players and coaches recommended a four-category variable after discussions and preliminary analyses. (Values: 1=22-24 ft., 2=24-25 ft., 3=25-26 ft., 4=26+ ft.; Original variable: Shot_Dist)
  7. Venue: Whether it was a home or away game for the shooter’s team. (Values: 0=Away, 1=Home; Original variable: Location)

Table 1 reports the logistic regression analysis results, notably, standard information such as logit coefficients, odds, z scores, and a measure of statistical significance (i.e., p>z).  It also reports useful non-standard information such as frequencies, (predicted) probabilities, and expected values.  The rationale for reporting standard and non-standard information, to borrow from the statistician Frederick Mosteller, is to “let weaknesses from one…be buttressed by strength from another” (16) (Ch. 4, p. 116), a concept he referred to as “balancing biases.”  As envisioned, data analysts can rely on standard information when building and evaluating logistic regression models, and non-standard when communicating the results and their implications to coaches and players.

Table 1.

Results of final logistic regression analysis

0-3 ft.6%1.290.86
3-6 ft.54%
6-9 ft.28%0.381.466.780.00.371.11
9+ ft.12%0.471.607.740.00.391.17
0-2 secs.5%1.210.62
2-4 secs.7%0.631.888.020.00.330.99
4+ secs.88%0.772.1711.870.00.361.08
Off catch75%1.361.07
Off dribble25%-
22-24 ft31%1.381.13
24-25 ft.36%-0.090.91-
25-26 ft.20%-0.170.84-
26+ ft.13%-0.300.74-

Note. n=32,511.  Log pseudolikelihood, starting value: -21,078.18; final value: -20,827.69.  Likelihood ratio (degrees of freedom=13): 498.44, p > chi2 = 0.00. Tjur R2: 0.014; McFadden R2: 0.012.  Stukel chi2(1) = 4.10, p > chi2 = 0.043

Standard versus Non-Standard Interpretations

Table 1 shows that the defender distance variable (DefDist) affects the outcome variable.  A standard interpretation would emphasize odds ratios and statistical significance:

Controlling for other variables’ effects, three-point shots taken with the closest defender 9+ feet away have a:

  • 60% higher odds (i.e., 1.6/1) of going in than those taken with the closest defender 0-3 feet away,
  • 24% higher odds (i.e., 1.6/1.29) than those with the defender 3-6 feet away, and
  • 10% higher odds (i.e., 1.6/1.46) than those with the defender 6-9 feet away.

Each effect is statistically significant, as their z scores show.

Although the standard interpretation is correct from a technical standpoint, coaches and players may not understand or act on it, given Zarren’s and Oliver’s comments (as well as those of DeMaris, Gelman, and Hill).  Now consider a non-standard interpretation (that relies on Table 1’s non-standard information).  Note that each percentage’s associated expected value is in parentheses.

All else unchanged, the percentage of three-point makes would decrease from 35% (1.05 pts.) to:

  • 29% (0.86 pts.) with the defender always 0-3 feet away from the shooter, and
  • 34% (1.02 pts.) with the defender always 3-6 feet away.

It would increase from 35% to:

  • 37% (1.11 pts.) with the defender always 6-9 feet away, and
  • 39% (1.17 pts.) with the defender always 9+ feet away.

NBA coaches and players would probably prefer the non-standard interpretation.  Arguably, reporting the likely effect in percentage points instead of odds is more intuitive and actionable (26, 30). 

Calculating Each Shot’s Make Probability

Another number to note in Table 1 is the constant of -1.46 logits which translates to a predicted make probability of 19% (0.56 pts.).  The -1.46 number represents a three-point shot with the lowest value on each predictor variable:

  • Defender 0-3 feet away
  • 0-2 seconds on the shot clock
  • Off the catch
  • First period
  • Shot distance of 22-24 feet
  • Away game

An implication is that it is possible to calculate the predicted make probability of each of the 32,511 shots.  Such information can spark curiosity and foster improved performance for a player scrutinizing his own (or opponents’) shot data.  For example, Row 1 of Table 2 reports the logit coefficients associated with the first three-point shot Klay Thompson of the Golden State Warriors attempted in 2014-15.  In the third period of an away game versus the Sacramento Kings with 4.6 seconds on the shot clock, Thompson missed from 22 feet off the catch with the defender 3.9 feet away.  As the column titled Prob shows, that shot’s predicted make probability was 38% (.38*100), calculated by applying the following formula to select Table 2 numbers: exp (sum of logit coefficients + constant)/ (exp (sum of logit coefficients + constant) +1).

Upon closer examination, Thompson could have asked the team’s data analysts how that shot’s make probability would have changed had the defender been 9+ rather than 3.9 feet away.  To respond, an analyst could have replaced the DefDist logit coefficient of 0.25 with 0.47, the one corresponding to a 9+ feet value.  As shown in Row 2, the make probability would have risen to 42%, a four-percentage-point increase or likely effect. 

Thompson next might have asked how shooting off the dribble rather than the catch would have affected the 42% probability.  After replacing the Catch logit coefficient of 0 with-0.09, an analyst could have reported that the probability would have dropped to 39%, as Row 3 of the Prob column shows. 

Thompson, an excellent shooter, would probably work to improve specific aspects of his shooting if he had such data for all his three-point shots (31).

Table 2.

Simulating the effect of changes on a single shot’s make probability 

Row DefDist ShotClock Catch Period ShotDist Venue Cost Total Prob 
0.25 0.77 -0.05 -1.46 -0.49 0.38 
0.47 0.77 -0.05 -1.46 -0.27 0.42 
0.47 0.77 -0.09 -0.05 -1.46 -0.36 0.39 

Predicting the Likely Effect of Multiple Changes to Multiple Predictor Variables

Coaches thinking more broadly might focus on all 32,511 shots and ask analysts to predict the likely effect of multiple changes to the values of multiple predictor variables. Building on the Thompson example, analysts could approach the task by conceptualizing changes as scenarios.  Below, and graphically in Figure 1, are three illustrative ones.

Scenario 1. Players take all 32,511 three-point shots with the defender 9+ ft. away.  

Prediction: 39% of all three-pointers will go in, an increase of four percentage points compared to the 35% baseline, translating to 1,297 more makes and 12,723 total ones.

Scenario 2. Players take all 32,511 three-point shots:

  • with the defender 9+ feet away 
  • from 22-24 ft. away from the basket

Prediction: 42% of all shots will go in, a three-percentage-point gain vs. Scenario 1.  This translates to 808 more makes and 13,531 total makes.

Scenario 3. Players take all 32,511 three-point shots:

  • with the defender 9+ ft. away 
  • from 22-24 ft. away from the basket
  • with 4+ seconds on the 24-second shot clock

Prediction: 43% of all shots will go in, an increase of another percentage point compared to Scenario 2, translating to 370 more makes and 13,901 total ones.

Figure 1.

Percentage of predicted makes by scenario 

Each scenario’s likely effect results from all-or-nothing simulation.  How does it work?  For any predictor variable, such as Catch, data analysts select one target value—either “Off Catch” (occurring 75% of the time) or “Off Dribble” (25%).  Assume they choose “Off Catch,” with a logit coefficient of 0, as Table 1 shows.  For the 8,127 “Off Dribble” shots, they would replace the coefficient of -0.09, also shown in Table 1, with 0 and calculate the new likely effect: 158 more made three-pointers for the season, translating to 11,584 total makes. 

Adopting a fine-tuning approach is another possibility.  After examining the frequency distribution of the Catch values, analysts could specify a new distribution, such as 92% “Off Catch” and 8% “Off Dribble,” ensuring the total sums to 100%.  They would keep the original 24,384 “Off Catch” values (i.e., 75%) and change the -0.09 coefficient to 0 for another 2,600 selected randomly from the original 8,127 “Off Dribble” values to achieve the 92:8 ratio.  The change would result in 11,530 made three-pointers, 54 less (i.e., 11,584-11,530) than if players had taken all shots off the catch.

If coaches and players embrace simulation, there could be too many scenarios for data analysts to handle.  To stay ahead of demand, they could build self-serve simulators tailored explicitly for coaches’ and players’ use.  Finding prototypes in academic research will be a struggle, however, arguably because of the non-linear relationship between logits and probabilities (26, 30) and its dampening effect on reporting likely effects in probabilities or percentage points.  Figure 2 plots illustrative logit and probability values to cast light on that relationship.

Figure 2.

The non-linear relationship between logits (x-axis) and probabilities (y-axis) 

Note how a one-logit increase from zero to one on the x-axis corresponds to a .23 probability increase (from .5 to .73) on the y-axis.  Yet a one-logit increase from four to five (or minus 5 to minus 4) translates only to a tiny probability increase.  As shown in Table 1 (and later in Table 3), it is still possible to report the effect of a predictor variable, x, on a binary outcome, y, in probabilities or percentage points (e.g., a one-unit change in x is associated with a three-percentage-point increase in y, all else being equal).  Arguably, it is also sensible to do so, not least because NBA players make roughly 35% of their three-point shots and the relationship between logits and probabilities is reasonably linear between .2 and .8 on the probability scale, as Figure 2 shows.  But in more extreme cases, as Figure 2 suggests, the effect size will depend heavily on the value of y and the values of the model’s other predictor variables.  More precisely, the size of the effect will decrease near 0 and 1.  As a result, x’s effect on y in probabilities percentage points “…cannot be fully represented by a single number” (19) (p. 23).  That may be why some logistic regression experts (6-8) have advised against using probabilities or percentage points to report and interpret logistic regression coefficients’ overall effects.  It also may be why most major statistical software packages do not produce effects in probabilities or percentage points through pre-packaged procedures or built-in modules.  As an unintended consequence, some data analysts seeking guidance likely have had to fend for themselves.           

Data analysts can use this guide to build simulators that report likely effects in probabilities or percentage points.  (For convenience, references are made to the three-point shot data used in this paper’s analyses, although the guide is general and should work across areas of interest.)  Several steps are involved in the process:

Step 1. Ensure sufficient three-point shot data are available to conduct logistic regression analysis, which should be a straightforward task for NBA teams given the league’s business relationship with Second Spectrum (which replaced SportVU).  How does one define sufficient?  As a rule of thumb, at least 10 shot attempts are needed for each predictor variable in logistic regression model, adjusting for the expected shot make rate (or miss rate if it is lower than the make rate).  For context, this paper’s main analysis with six predictor variables and a 35% expected make rate required a minimum of 171 three-point shot attempts: 10 * (6 /.35).  For non-NBA teams requiring raw data, assistant coaches can record key shot characteristics with paper and pencil or specialized hand-held apps. 

Step 2. Develop a model to predict successful 3-point shots, the binary outcome of interest.  Logistic regression produces a weight—a logit coefficient—for each category of each predictor variable.  In an optimal model, those weights maximize the predicted probability gap between the mutually exclusive outcomes (1).  

Step 3. To calculate a single 2014-15 three-point shot’s make probability, sum the weights corresponding to its characteristics and add the constant.  After that, apply the formula shown earlier to the result: exp (sum of logit coefficients + constant)/ (exp (sum of logit coefficients + constant) +1).  Alternatively, request the predicted probability from the statistical software.

Step 4. Do the same for the 32,510 remaining shots, sum all 32,511 probabilities, then take the average to compute the overall make probability.  If the model predicts players will make 35% of all three-point shots, it translates to 11,426 makes (.35*32,511).   

Step 5. To enable the simulator to work online or in a mobile app, develop an algorithm using JavaScript.  The simulator’s purpose is to let users see how changes they make to the values of the predictor variables affect the .35 probability.  

Step 6. Design a user interface, possibly by enlisting the support of someone familiar with website and app development.

Step 7. Keep things simple initially—permit users to change only one value of one predictor variable.  If it has two response choices like Away and Home, let the user change every Away response to Home or vice versa.  Think of this as the all-or-nothing option.  

Step 8. For all 32,511 three-point shots, change the corresponding Away or Home logit coefficient (but no others) to align with the user’s selection, then recalculate the predicted make probability.  The likely effect is the difference between the new and starting probability (and the new and starting makes).   

Step 9. Follow the same process to let users change the values of several predictor variables simultaneously. 

Step 10. Go further and allow users to change any predictor variable’s frequency distribution as they please, ensuring the distribution sums to 100%.  Think of this as the fine-tuning option.  The algorithm will need rules to accommodate the changes.  

What would all-or-nothing and fine-tuning self-serve simulators look like, and how would they function?  Figure 3 shows a screenshot of a working all-or-nothing simulator (accessible at  The first column contains the predictor variables and their values.  Column 2 shows the changes (in blue) the user made to the 2014-15 frequencies; the third column displays the original frequencies.

Figure 3

All-or-nothing simulation 

As Figure 3 shows, the user selected values of “0-3 ft.” for “Defender Distance,” “0-2 secs.” for “Time Left on Shot Clock,” “Dribble” for “Off Catch or Dribble?” and “26+ ft.” for “Shot Distance.”  The likely effect is a 22-point decrease in the make probability, translating to 7,229 fewer makes and 4,197 total ones.

Personalized simulators for players like Klay Thompson and Stephen Curry could be more beneficial (and accurate) than a generic, all-player one.  To support this point, Table 3 reports the results of a new analysis of Curry’s 2014-15 three-point shots.  Note how the values of many key measures, such as frequencies and expected values, differ substantially from their Table 1 counterparts.  Table 3 shows, for instance, that Curry took 54% of his three-pointers off the dribble with an expected value of 1.32 points per shot.  But Table 1 showed NBA players (including Curry) took only 25% of their three-pointers off the dribble with a 1.01 points-per-shot expected value.  Curry is not your average three-point shooter, hence the need for personalization.  

Table 3.

Results of Steph Curry logistic regression analysis 

0-3 ft.11%1.240.72
3-6 ft.55%0.892.442.440.02.431.29
6-9 ft.24%0.972.652.480.01.451.35
9+ ft.10%1.263.512.750.00.521.55
0-2 secs.2%1.250.75
2-4 secs.3%
4+ secs.95%0.792.
Off catch46%1 .401.21
Off dribble54%
22-24 ft16%1.551.65
24-25 ft.31%-0.750.47-2.460.01.371.11
25-26 ft.24%-0.510.60-1.580.11.431.28
26+ ft.28%-0.650.52-

Note.  n=j.  Log pseudolikelihood, starting value: -305.04; final value: -294.46.  Likelihood ratio (degrees of freedom=13): 21.16, p > chi2 = 0.07. Tjur R2: 0.047; McFadden R2: 0.035.  Stukel chi2(1) = 4.38, p > chi2 = 0.11.

A working fine-tuning simulator—a complement to the Curry analysis—is available at  It lets users change any value of any predictor variable by any amount and see the likely effect.  In the screenshot shown in Figure 4, the user changed Curry’s 2014-15 season frequencies (in parentheses) for “Defender Distance,” “Off Catch or Dribble?” and “Shot Distance.”  The likely effect is a seven-percentage-point increase to his 42% average make probability, translating to 31 more makes (i.e., 220-189).

Figure 4 

Steph Curry’s fine-tuning simulator 


If the sample size of three-point shots allows, data analysts can build all-or-nothing and fine-tuning simulators that include all teams and players, each team, and each player.  Given sufficient demand, they can also do so with data for other major shot types (i.e., two-pointers and free throws).    

Several caveats are in order before describing how basketball teams might act on the results the approach described here, using the results (and simulators) shown earlier for illustration.  First, inferences drawn from the 2014-15 dataset may no longer apply because of the time gap.  Nor did this dataset include several three-point shot characteristics (e.g., closest defender’s height and reach, the game score at each shot) that could be important, which is a second caveat. 

A third caveat concerns the “all else the same” assumption, a logistic regression analysis theoretical staple.  In practice, it may not hold up.  Giving excellent three-point shooters more playing time, for example, could worsen teams defensively.  Deciding who plays and why, a type of optimization, lies outside this paper’s scope.

Another caveat involves ease of implementation.  Building and updating simulators like Curry’s for NBA players who shoot, say, 175 or more three-point shots per season may require automation.  To characterize the task as trivial would be misleading.

Humility and ignorance are two key factors to consider as the fifth caveat.  Some NBA data analysts may have already adopted an approach combining good data, logistic regression, likely effects reporting in probabilities or percentage points, and self-serve simulation.  As noted earlier, they work mainly in secrecy.  And when they make comments at analytics conferences or similar forums, some are instructed “to go up on stage and talk about something without saying anything” (15) (51:37), according to Zarren.

Application In Sports

Good basketball coaches position their players to make the highest percentage of three-pointers possible, all else equal.  They also implement a defense to minimize opponents’ three-point make percentage.  The analyses presented here suggest six factors affect the make percentage:

  • Closest defender’s distance to the shooter
  • Time left on the 24-second shot clock
  • Whether the player shot off the dribble or catch
  • Game period
  • Shot distance
  • Venue

How might coaches act on these findings?  There are numerous possibilities, starting with game pace.  Fast ball movement from defense to offense (e.g., before the defense sets) gives the offensive team more time to find an open three-point shot, preferably before the four-second mark on the shot clock where shooting percentages dip, and unquestionably before the two-second mark where they plummet.  As the NBA’s all-time leading three-point shooter, Steph Curry understands this well.  Table 3 showed he attempted only two percent (compared to a five percent NBA average) of his three-point shots with less than two seconds on the shot clock.

Coaches should design offensive plays and patterns to create at least three feet of space between the shooter and defender.  A 22-24-foot shot’s make probability with the defender 0-3 feet away is only 29%, all else equal.  It increases to 34% with the defender 3-6 feet away.  Space is critical for Curry, too.  He shot 11% of his three-pointers with the defender 0-3 feet away versus the NBA average of 6%, reducing his overall make percentage.  It could have been worse.  Had he taken all 448 of his shots with the defender 0-3 feet away, all other factors being equal, his make probability would have dropped from 42% to 24%.

Making sure players understand the characteristics of a desirable three-point shot is another opportunity.  Personalized simulators like Curry’s can make each player’s shooting strengths and weaknesses obvious.  For instance, some players may make a higher percentage of three-pointers off the dribble than catch.  Others may suffer only a slight percentage point decline when guarded tightly or shooting from 26+ rather than 22-24 feet.  And if those simulators contain opponents’ shot data, coaches could use them to determine how to exploit specific opponents’ weaknesses.

Analyses show the three-point make percentage drops in the fourth period.  Player fitness could be a contributing factor.  Without applicable data (e.g., feet, meters, or miles logged since tip-off), it is difficult or impossible to test the hypothesis.  Maybe the players on the court lack the skills needed to shoot higher percentages.  Or game stress could affect shooting performance—data on the game score at each shot would clarify the matter.  For context, the all-or-nothing simulator would show that the highest probability three-point shot (46%) has these characteristics:

  • Defender 9+ feet away
  • 4+ seconds on the shot clock
  • Off the catch
  • First period
  • 22-24 feet from the basket
  • At home 

The simulator would also show that the 46% make probability drops to 42% in the fourth period, changing nothing else.  That means players have grown tired, different players are on the court, game pressure has taken its toll, or unknown variables caused the drop.  So how should head coaches make sense of this?  Working with assistant coaches and data analysts, they can explore ways to increase players’ fitness levels, optimize substitution patterns, and help players cope better with pressure.  If teams can access variables that were unavailable for analysis here, their analysts can include them in new models to estimate their likely effect.

Players make a higher percentage of three-point shots at home than on the road, all else equal.  Crowd noise, characteristics (e.g., lighting) of the less familiar setting, travel effects (e.g., uncomfortable hotel beds), or some combination of these may explain why.  Coaches can look outside the league for ideas to help players overcome such obstacles.  For instance, former US Navy SEAL commander Mark Divine prepares SEAL candidates for training by replicating the challenges they are likely to encounter, including Hell Week during which “each candidate sleeps only about four total hours but runs more than 200 miles and does physical training for more than 20 hours per day” (5). 

Contrary to conventional wisdom, Divine’s SEALFIT program places particular emphasis on skills like positive visualization, breath control, and meditation because, as he said, “People who haven’t learned to control their mind and emotions quit or they get hurt” (10).  Does SEALFIT work?  Divine reports that nine of 10 SEAL candidates who complete SEALFIT training become SEALs (versus a 20% norm).  He is confident that NBA players would benefit from the program (M. Divine, personal communication, March 11, 2022).

A complementary tool for improving performance is psychotherapy.  As described earlier, Ben Simmons’s decision to avoid attempting an open lay-up or dunk (arguably) for fear of being fouled and having to shoot free throws may have cost his team the 76ers a 2021 playoff series to the Hawks.  As his teammate Joel Embiid declared, “That was the turning point” (12) (1:08).  Psychotherapist Richard Schwartz, who developed the Internal Family Systems (IFS) therapeutic model (23), would probably concur then speculate that Simmons’s widely criticized decision (21, 27) originated from past trauma linked to his poor free-throw shooting.  After citing evidence (24) of IFS’s effectiveness, Schwartz might posit that a protective part of Simmons’s mind—a “guardian of [his] inner world” (23) (p. 184)—compelled him to pass rather than shoot to prevent a traumatized part—think of it as a deeply wounded child—from re-experiencing pain or shame at the free throw line.  Were Schwartz to work with Simmons, he would likely try to communicate with his mind’s traumatized part as if it were an actual person, restore its faith in Simmons’s free-throw shooting abilities, and encourage the protective part to undertake different tasks.  The more traditional coaching approach of advising, or even requiring, Simmons to practice harder with expert guidance did not—and may never—work.  As Early (9) observed, “Simmons has been reluctant to seek help from top shooting coaches…He reportedly clashed with his former team (the 76ers) years ago over who he would work with, preferring to practice with his brother rather than team shooting coach John Townsend.” 

Coaches can use the same strategies to reduce their opponent’s three-point shooting percentage they use to improve their own.  Table 1 data (and the all-or-nothing simulator) suggest the key lies in forcing opponents to shoot with less than four seconds on the clock, off the dribble, from long distances while being closely guarded.  Stepping up the defensive intensity in the first and third periods where the likelihood of making a three-point shot is relatively high, and motivating the home crowd to unsettle opponents makes sense, too.

Coaches can also think about implementing a full- or three-quarter court press more often, maybe for entire games.  The goals of a 2-2-1 three-quarter court press, for example, are control and containment, not turnover generation.  As envisioned, its use would slow down the game and force opponents to shoot a higher percentage of difficult three-pointers with less time on the clock, reducing their make percentage.  As Hall-of-Fame coach Jack Ramsay explained in Pressure Basketball, “The tempo of the game is controlled by the defensive team and the best manner of control is through the exertion of pressure at some point on the court” (22) (p. 80).

Good data, logistic regression analysis, and self-serve simulation can also promote truth and trust, positive attributes for any coach or leader.  Maybe tongue in cheek, former NBA coach Jeff Van Gundy (15) (17:40) confessed to lying to his players. “If I saw what I wanted to change,” he said, “I would either use numbers to support it or make them up because the players are not going to know the difference.”  Giving players tools that predict the likely effects of their potential actions would be more truthful and potentially more effective, too. 


Keeping things simple is critical in basketball.  According to Zarren (15) (7:00), “There are 20 things in (the coach’s) head that will get us X number of wins per season, but you can only focus on six of them in practice, and the players might only remember four and actually execute one in a game.  So you’ve got to pick your battles if you’re a stats guy who…needs to talk to a coach.  But if you’re a coach, you need to pick your battles, too.”

Van Gundy (15) (16:51) offered data analysts and coaches strong advice related to this point from his coaching experience.  “I wouldn’t tell a guy you’re 38% on three to four dribbles so dribble a fifth time because you go up to 40%,” he said.  “You better be pretty sure about what you’re saying…You want players to feel confident.  You don’t want them out there saying, ‘Was that [four] dribbles or [five] when I pull up?’” 

To mitigate the risk of generating harmful insights, data analysts should actively engage coaches and players in making key analytical decisions (e.g., ensuring predictor variables and their levels are meaningful), not least because Van Gundy and others who share his philosophy consider basketball sense—the capacity to make wise choices that benefit the team—to be of paramount importance.  

Arguably, self-serve simulation with likely effects reporting in probabilities or percentage points is steeped in such basketball sense.  As a benefit, data analysts will not need to rely on technical terms (e.g., “he shoots two standard deviations below the league average when you force him to the left” (15) (48:20)), as former Memphis Grizzlies’ executive John Hollinger once did.  Instead, they can speak with more authority using plain language (e.g., “his make probability drops to 28% when you force him to the left”).  Or they can make self-serve simulators available to players (and coaches) and let them figure it out on their own.  They may appreciate it, even cynics sharing Hall-of-Fame player Charles Barkley’s views: “Analytics don’t work at all.  It’s just the crap that some people who are really smart made up to try to get in the game because they had no talent” (29) (2:05).

NBA and other basketball teams worldwide should consider adopting an approach that combines good data, logistic regression analysis, likely effects reporting in probabilities or percentage points, and self-serve simulation.  The possible benefits are myriad.  It can help teams increase their three-point shooting percentages while lowering their opponents’; improve communication among data analysts, coaches, and players; enhance each group’s effectiveness; and lead to more wins. 


Variables in the 2014-15 NBA shot dataset

  1. Game_Id: The game’s unique identifier.
  2. Matchup: The teams competing.
  3. Location: Whether it was a home or away game for the shooter’s team.
  4. Outcome: Whether the shooter’s team won or lost.
  5. Final_Margin: By how many points the shooter’s team won or lost.
  6. Shot_Number: The shooter’s nth shot that game.
  7. Period: The period in which the shooter took the shot.
  8. Game_Clock: Minutes and seconds left in the period in which the shooter took the shot.
  9. Shot_Clock: Seconds remaining on the 24-second shot clock when the shooter took the shot.
  10. Dribbles: Number of dribbles the shooter took before shooting.
  11. Touch_Time: Number of seconds the shooter had the ball before shooting.
  12. Shot_Dist: Distance in feet from the center of the basket to the shooter.
  13. Pts_Type: Whether the shooter attempted a two- or three-point shot.
  14. Shot_Result: Whether the shooter made the shot.
  15. Closest Defender: Name of the defender closest to the shooter.
  16. Closest_Defender_Player_Id: The closest defender’s unique identifier.
  17. Close_Def_Dist: The closest defender’s distance to the shooter in feet.
  18. Fgm: Whether the shooter made the shot.
  19. Pts: The shot’s point value (0, 2 or 3).
  20. Player_Name: The shooter’s first and last name.
  21. Player_Id: The shooter’s unique identifier.

Note: The original dataset contained 128,069 two- and three-point shots. After removing all two-point shots, and all three-point shots with a missing (or unimputable) value on the Shot_Clock variable, the size decreased to 32,511. For a value to be imputable, there had to be 24 seconds or less on the game clock when the player took the shot. In that case, the decision was made to replace the missing Shot_Clock value with the Game_Clock value.


The author would like to thank David Clemm, Robert Eisinger, Ward Fonrose, John Geraci, Ryan Heaton, Adam Hoeflich, Priam Lacassagne, Roxane Lacassagne, and Mark Naples for reviewing earlier versions of this paper, and for providing helpful comments and suggestions. The author is particularly thankful to Dan Dougherty (who passed away in 2022) and Tom Northrup for their indirect contribution. Their longstanding beliefs and ideas about how basketball should be played permeate this paper’s “implications for coaches” section.


  1. Allison, P. (2013, February 13). What’s the Best R-Squared for Logistic Regression? Statistical Horizons.
  2. Allison, P. (2015, April 1). What’s So Special About Logit? Statistical Horizons.
  3. Basketball Reference. (2023).
  4. Ben Simmons passes up a wide-open dunk Sixers vs Hawks Game 7. (2021, June 20).
  5. BUD/S Hell Week. (2015, February 25). Navy SEALs.
  6. DeMaris, A. (1992). Logit modeling: practical applications. Sage Publications.
  7. DeMaris, A. (1993). Odds versus Probabilities in Logit Equations: A Reply to Roncek. Social Forces, 71(3), 1057-1065.
  8. DeMaris, A.; Teachman, J.; Morgan, S. P. (1990). Interpreting Logistic Regression Results: A Critical Commentary. Journal of Marriage and the Family, 52(1), 271-277.
  9. Early, D. (2022, February 24). Ben Simmons Savagely Roasted by Legendary Philly “Shot Doctor.” ClutchPoints.
  10. Eighty Percent of Navy SEAL Candidates Fail for a Reason. (2017, September 14). SEALFIT.
  11. Gelman, A. B., & Hill, J. (2009). Data analysis using regression and multilevel/hierarchical models. Cambridge University Press.
  12. Joel Embiid blames Ben Simmons for game 7 loss…. (2021, June 20).
  13. Johnson, E. [@MagicJohnson]. (2021, June 16). Give Hawks coach Nate McMillan a lot of credit he did the hack-a-Shaq on Ben Simmons to send him to the free throw. [Tweet]. Twitter.
  14. Meehan, B. (2017). Predicting NBA Shots.
  15. MIT SLOAN Analytics Conference: Basketball Analytics. (2012, March 12) Retrieved November 20, 2023, from
  16. Mosteller, F. M. (1996). Discussant comments for So what? The implications of new analytic methods for designing NCES surveys by Robert F. Boruch and George Terhanian. In From Data to Information: New Directions for the National Center for Education Statistics, Hoachlander, G.; Griffith, J.E.; Ralph, J.H.; US Department of Education, National Center for Education Statistics: NCES 96–901, pp. 4-116-4-118.
  17. NBA shot logs. (2016).
  18. Nourayi, M; Singhvi, M. (2021, January 15). The Impact of NBA New Rules on Games. The Sport Journal.
  19. Pampel, F. C. (2000). Logistic Regression. SAGE Publications.
  20. Peterson, D. (2020, May 28). How Different Metrics Correlate with Winning in the NBA over 30 Years. Medium.
  21. Pina, M. (2021, June 20). Ben Simmons’s Flaws Laid Bare in Potential End of the Process. Sports Illustrated.
  22. Ramsay, J. (1963). Pressure Basketball.
  23. Schwartz, R. C. (2023). Introduction to Internal Family Systems therapy (2nd ed.). Sounds True.
  24. Shadick, N. A.; Sowell, N. F.; Frits, M. L.; Hoffman, S. M.; Hartz, S. A.; Booth, F. D.; Sweezy, M.; Rogers, P. R.; Dubin, R. L.; Atkinson, J. C.; Friedman, A. L.; Augusto, F.; Iannaccone, C. K.; Fossel, A. H.; Quinn, G.; Cui, J.; Losina, E.; Schwartz, R. C. (2013). A Randomized Controlled Trial of an Internal Family Systems-based Psychotherapeutic Intervention on Outcomes in Rheumatoid Arthritis: A Proof-of-Concept Study. The Journal of Rheumatology.
  25. Stats LLC and NBA to make STATS SportVU Player Tracking data available to more fans than ever before. (2016, January 19). NBA Communications.
  26. Terhanian, G. (2019). The Possible Benefits of Reporting Percentage Point Effects. International Journal of Market Research, 61(6), 635–650.
  27. Thomas, L. (2021, October 3). Ben Simmons and the Acceptance of Failure. The New Yorker.
  28. Thorp, E. O. (2018). A man for all markets: from Las Vegas to Wall Street, how I beat the dealer and the market. Random House.
  29. TNT’s Charles Barkley rants about analytics in NBA, Houston Rockets GM Daryl Morey. (2015, February 10).
  30. Williams, R. (2012). Using the Margins Command to Estimate and Interpret Adjusted Predictions and Marginal Effects. The Stata Journal, 12(2), 308–331.
  31. Zwerling, J. (2014, August 27). Team USA’s Klay Thompson Breaks Down the Skills That Make Him a Shooting Star. Bleacher Report.
2024-05-21T13:46:56-05:00May 17th, 2024|General, Research, Sports Management|Comments Off on Advice on making the most of basketball three-point shot data
Go to Top