You are on page 1of 10

Original research

International Journal of Sports Science


& Coaching
Factors associated with coach–athlete 2021, Vol. 16(3) 509–518
! The Author(s) 2021
conversations about mental health in Article reuse guidelines:
sagepub.com/journals-permissions
intercollegiate sport DOI: 10.1177/1747954121993455
journals.sagepub.com/home/spo

Jessica Murphy and Philip Sullivan

Abstract
Increasing conversation about mental health is one way to decrease stigma and prevent/treat maladaptive processes
within the context of sport. Due to their proximity and influence over athletes, coaches can increase conversation and
change team culture. The objective of this study was to apply the Theory of Planned Behavior (TPB) to explore the
factors associated with coach-athlete conversation about mental health. A total of 136 Canadian coaches completed a
demographic questionnaire as well as a TPB-based survey measuring Attitudes, Perceived Behavioral Control (PBC), Social
Norms and Intention. Intention was measured as Role perception, if a coach believed it was their role to be involved in
athlete mental health. Behavior was measured as talking with an athlete(s) about mental health. Approximately 68% of
coaches had spoken to athletes about mental health in the last season. The linear regression model predicted a signif-
icant amount (42.7%) of the variance in Intention (p < .05) from the three TPB constructs. Logistic regression found a
significant interaction effect of PBC and Intention on Behavior (p < .01). Measured TPB construct scores were influenced
by previous mental health training, personal experience with mental illness, age group and the act of talking (p < .05).
Although a promising amount of coaches spoke to athletes about mental health, improvement is still possible. Mental
health training should continue to be promoted to all members of the athletic community. As attitude scores were
generally positive, this training should potentially focus more on improving capabilities and social norms.

Keywords
Attitudes, coach behavior, social norms, theory of planned behavior

Introduction
with mental illness.4 A 2019 study by Sullivan and col-
Student–athletes are a unique group as they belong to leagues7 reported that 20% of surveyed student–ath-
three separate cohorts, all identified as high risk for letes showed psychological distress scores indicative
mental health disorders. Firstly, student–athletes fit of severe mental distress. The distress scores of stu-
into the classification of “emerging adults,” those dent–athletes were significantly higher than their non-
between the ages of 18 and 24 years, who have the high- athlete peers and the general population.7 In addition
est prevalence of mental illness of all age groups.1,2 to psychological distress, many mental illnesses have
Further, student–athletes are enrolled in post- been identified as highly prevalent within the athletic
secondary education; students are two-times more population.8–10 Anxiety and depression are among the
likely to report symptoms of mental illness and elevated most common mental health disorders with approxi-
distress as compared to their non-student peers.3 mately 30% of student–athletes reporting insurmount-
Lastly, student–athletes compete in high-level sport, able difficulties during the past month and almost 25%
which introduces specific stressors including symptoms
of overtraining, injury, competitive failure, organiza-
tional factors, excessive time commitments, weight Reviewer: Britt Brewer (Springfield College, MA, USA)
control, and maintaining “appropriate” athlete behav- Department of Kinesiology, Brock University, St Catharines, ON, Canada
iour.4–6 Such additional pressures appear to override Corresponding author:
the benefits of physical activity and make student–ath- Jessica Murphy, 1812 Sir Isaac Brock Way, St. Catharines, ON, Canada.
letes at least as likely as non-athlete students to struggle Email: jm09ah@brocku.ca
510 International Journal of Sports Science & Coaching 16(3)

reporting clinically relevant levels of depressive desired outcomes and/or rewards, and when the actions
symptoms.11,12 are expected or required within their work role.21
Unfortunately, within the athletic community, in As earlier research has outlined the importance of
combination with the high risk of mental illness, open discussion and conversation about mental health
there is a lack of willingness to seek help, low levels within the athletic environment, it is important to
of perceived social support, and high levels of stigma- explore what factors may contribute to the performance
tization.5,8–10 According to surveyed athletes, many of of this behavior. Understanding such factors is critical
the commonly listed barriers to positive mental health for creating a positive sporting environment, as well as
and help-seeking (i.e.: lack of mental health literacy in developing mental health-related interventions and
sport, stigma from coaches/teammates, and lack of resources for coaches. As outlined by Mazzer and
time) originate from the athletic environment.5,8,13 As Rickwood,21 it is possible that perceived outcomes and
such, researchers have suggested that team culture capabilities, as well as role expectations may impact
needs to be improved.5,9 Due to the influence, proxim- whether a coach performs helpful behaviors. One
ity, and frequency of time that coaches have with ath- theory that could be used to explore this relationship is
letes, they are often chosen as avenues to improve team the Theory of Planned Behavior.24 According to this
culture.14–16 Consensus statements on the mental theory, an individual’s behavior can be predicted
health of athletes have urged coaches to “be aware of through their Intention to perform said behavior. That
the person behind the player” by being open to discus- Intention is influenced by three constructs, Attitudes,
sions about mental health.14,16 General sport psychol- Perceived Behavioral Control (PBC) towards the behav-
ogy research shows that student–athletes feel supported ior and the perceived Social Norms of others.24 Attitudes
when there is open and honest communication, and refers to one’s beliefs or thoughts towards the behavior
when the sporting staff is present and available.17 and the outcome that performing the behavior may
Social support not only helps with positive physical, create. Perceived Behavioral Control (PBC) focuses on
mental, and emotional development but also lessens the individual’s ability to perform the behavior. Lastly,
the load of various stressors endured by student– Social Norms refers to the perceived beliefs of important
athletes.17 This sentiment is mirrored by mental others towards the behavior.24,25
health-specific research in elite and student–athletes The TPB, or variations of the TPB, have been applied
with researchers concluding that having more conver- in many research areas, including both mental health and
sations about mental health can help to prevent and sport. For example, the TPB has been used to measure
effectively treat maladaptive processes.5,16,18,19 predictors of faculty intentions to refer26 and respond
Fortunately, coaches at the recreational, competitive to27 students believed to be struggling with a mental
and intercollegiate levels agree that it is within their health concern. It has also been used with community
role to be involved in athlete mental health.20–22 youth workers and coaches to determine factors associ-
While encouraging, evidence suggests that further ated with the intent to advise youth to seek professional
investigation into this area is warranted. Although help.28 In addition, the TPB model has been used in a
coaches agreed they should be involved in mental variety of sport settings to predict concussion reporting
health, their behaviors do not completely match this and evaluate the effects of education on such report-
opinion. Mazzer and Rickwood21,23 found that most ing.29–32 Although the methods used to measure the
coaches performed “promotional” behaviors (e.g.: pro- TPB constructs are self-reported, and therefore may
moting living a healthy lifestyle, identifying and stop- influence biases, results are still useful in understanding
ping bullying), as opposed to “prevention” or behavior and guiding intervention development. The
“intervening” behaviors. The least commonly engaged purpose of this study was to use the TPB to determine
activities included asking about suicide risk, talking which, if any, of the three constructs (Attitude, PBC or
about early warning signs of mental illness, offering Social Norms) were associated with coach self-reported
advice about mental health concerns, and contacting intention and, further, if intention was associated with
parents or professionals about the struggling individu- coach self-reported behavior. The secondary objective of
al.23 These results suggest that, although most coaches this study was to explore demographic factors that may
state they believe it is within their role to support ath- be associated with the three constructs and the Behavior.
lete mental health, it may be that some coaches are not
taking the proper actions to do so. Mazzer and
Rickwood21 proposed that coaches may need further Methods
and additional motivators to perform helpful behav-
Procedure
iors. They hypothesized that coaches may be more
likely to perform helpful behaviors when they perceive Canadian coaches were contacted through either uni-
themselves as capable, when the behavior is linked to versity athletic directors, or the Coaching Association
Murphy and Sullivan 511

of Canada. Athletic directors of Canadian universities Table 1. Demographic variables of participating coaches.
were contacted via their publicly available email
Demographic variable n (%)
address. Emails contained a brief description of the
study along with a link to a consent form and ques- Gender
tionnaire. The athletic directors were asked to circulate Male 83 (61.0%)
this invitation to all coaches at their institution. The Female 50 (36.8%)
Non-Binary 2 (1.5%)
Coaching Association of Canada posted the brief
Missing 1 (0.7%)
description and the link to the questionnaire on their
Position
coach’s portal. Interested coaches who followed the Head Coach 84 (61.8%)
link were first taken to a consent form where they pro- Assistant Coach 47 (34.6%)
vided voluntary consent; they were then forwarded to Prefer not to Say 4 (2.9%)
the questionnaire. All coaches over the age of 18 were Missing 1 (0.7%)
eligible to participate. Ethnicity
Caucasian 119 (87.5%)
Other 9 (6.6%)
Participants
Asian 3 (2.2%)
A total of 136 coaches completed the study; demo- Black 3 (2.2%)
graphic data can be found in Table 1. The majority Latino 1 (0.7%)
were male (61%), head coaches (61.8%) and from Aboriginal 1 (0.7%)
Ontario (41.9%). Twenty-six different sports were rep- Personal experience
Yes 101 (74.3%)
resented in the current study; the most represented
No 34 (25.0%)
sports were hockey (18.7%) and track and field/
Prefer not to say 1 (0.7%)
cross-country (14.2%). Approximately 70% of coaches Previous training
strongly agreed or agreed that it was within their role as No 64 (46.3%)
a coach to be involved in athlete mental health. Yes 63 (47.1%)
A similar percentage (68%) reported that they had I do not Know 9 (6.6%)
spoken to their athlete(s) about mental health in the Age
last season. 40 years or younger 71 (52.2%)
41 years or older 63 (46.3)
Missing 2 (1.5%)
Data collection Competitive level
Both consent and data collection were completed Intercollegiate 35 (25.7%)
through Qualtrics,33 a secure online data collection Provincial Competitive 34 (25.0%)
platform. The questionnaire included demographic National Competitive 31 (22.8%)
variables (e.g.: age and gender) and experience- International Competitive 14 (10.3%)
Local Competitive 12 (8.8%)
specific measures (e.g. coaching role, previous mental
Non-Competitive/Recreational 8 (5.9%)
health training, personal experience with mental ill- Prefer not to Say 2 (1.5%)
ness). Prespecified age groups (18–25, 26–30, 31–35, Talked to athlete
36–40 and 40þ) were used to measure age in an attempt Yes 93 (68.4%)
to promote anonymity. Following data collection, it No 37 (27.2%)
was found that the number of coaches in each group I do not know 6 (4.4%)
was imbalanced. In response, the five age groups were n: number of participants; %: percentage of total sample.
collapsed into two (39 and 40 years) groups. This
classification is justified given earlier research stating
differences in mental health opinion and knowledge autonomy and capacity to perform the behavior
between those under and over 40.34–36 including questions on confidence, and perceived deter-
A TPB survey, developed following the guidance of rents the behavior (“athletes don’t need their coach to
Ajzen,24 was also completed by each participant. Five have positive mental health”). Social Norms questions
questions measured Attitude, seven measured PBC and asked if the coach though important others (fellow
two measured Social Norms. Attitude measures cap- coaches and athletes) would support the behavior.
tured the affective feelings and attitudes towards the Attitude, PBC and Social Norm questions were
outcome of performing the behavior. Attitude ques- answered using a 7-point Likert-Scale with 1 represent-
tions related to items like comfort, importance, and ing strongly agree and 7 representing strongly disagree.
level of benefit from talking to an athlete(s) about Intention was measured through asking coaches if they
mental health. Questions to measure PBC focused on believed that it was their role as a coach to talk to
512 International Journal of Sports Science & Coaching 16(3)

athletes about mental health. Self-identity and role additional model, outside of the TPB was also per-
were first introduced as an addition to the TPB formed to measure any possible interaction effect of
model, however, Ajzen and Fishbein37 believe that Intention and PBC on conversation with one or more
role, instead, is better as an alternative measure for than one athlete. Binary logistic regression was also
Intention.38 The Intention question was also answered used, with the interaction between Intention and PBC
on a 7-point Likert-Scale. Given the formation of the as the predictor and “number of athletes” as the depen-
Likert-Scale, higher scores on the three constructs and dent variable, with 0 representing “one athlete” and 1
on Intention, indicate less supportive/less positive representing “more than one athlete.”
Attitudes, PBC, Social Norm and Intention scores. Chi-Square tests were used to explore the associa-
Behavior was measured using the question “In our tion of both demographic and experience variables with
last season, I had a conversation about mental health Intention, Behavior and Conversation with one or more
with an athlete(s); this question was answered as than one athlete. Independent t-tests were used to deter-
“True,” “False,” or “I do not know/remember.” mine significant differences in construct scores between
Those coaches who indicated “True” were asked if to demographic and experiential variables and Behavior.
indicate if the conversation was with “one athlete” or During data analysis, only those participants who gave
“more than one athlete.” definitive answers were included, in other words, those
who responded with “I do not know/remember” or
Data analysis “prefer not to say” were excluded from analyses involv-
ing those variables. In addition, given the small number
For each participant, the raw questionnaire data were of coaches who identified as non-binary (n ¼ 2), these
used to create an average score for each construct. For individuals were also not entered into gender analyses
example, the five questions measuring Attitude were on TPB constructs. All analyses were done in IBM
averaged to create an average Attitude score for each SPSS Version 26.40 Graphical representations of the
participant. Calculated scores were then used in two data (Figure 2) were created using RStudio.41
separate regression models to determine the fit of the
TPB model to the data. A linear regression analysis
measured the ability of the three TPB constructs Results
(Attitude, PBC and Social Norms) to predict The Cronbach’s alpha for the entire survey was calcu-
Intention. As suggested by Hankins and colleagues,39 lated at 0.80. Individual constructs were lower with
for TPB data and regression analysis, the Adjusted R2 Attitude (5 Items) calculated at 0.65, PBC (7 Items)
value was used. A binary logistic regression model was at 0.72 and Social Norms (2 items) at 0.51.
then used to determine the interaction effect of PBC Acceptable Cronbach alpha values range from 0.70–
and Intention to predict. The interaction between PBC 0.95. A low Cronbach’s alpha value can be caused by
and Intention was added as the predictor, Behavior was a small number of questions, poor inter-relatedness
added as the dependent variable, with 0 representing between items of heterogenous constructs.42 Prior to
“Did not Talk” and 1 representing “Talked. The pro- conducting the analyses, the data was checked for
posed TPB model can be found in Figure 1. An assumptions of linear and logistic regression. For

(a) (b)

Atude

Social Norms Intenon Behavior

PBC

Legend: PBC, Perceived Behavioral Control

Figure 1. The theory of planned behavior model.


PBC: perceived behavioral control.
Murphy and Sullivan 513

linear regression, the Durbin Watson statistic (2.10) Social Norm scores increase (become less supportive)
suggested independence of observations. The the Intention score also increases (becomes less support-
Tolerance (TIF; range: 0.741–0.863) and the Variance ive of the belief that the role of a coach is to be involved
Inflation Faction (VIF; range: 1.159–1.349) statistics in athlete mental health) (Table 2). The binary logistic
suggested no multicollinearity. Skewness and Kurtosis regression model revealed a good fit of the model to the
z-scores were calculated by dividing the Skewness and data (X2(1) ¼ 5.647, p < .05). Results indicated a signif-
Kurtosis values by the standard errors; all were within icant negative association between the interaction of
the acceptable range (z <  3.29).43 PBC and Intention on Behavior (Table 2; Figure 2).
Approximately 70% of coaches strongly agreed or Those coaches with a lower interaction PBC and
agreed that it was within their role as a coach to be Intention score tended to be more likely to report
involved in athlete mental health. There were no signif- having performed the Behavior than those with a
icant associations found between any of the demo- higher interaction score. Attitude, PBC and Social
graphic variables and Intention (p > .05). A significant Norms were all significantly correlated with each
association between Intention and previous mental other, as hypothesized in the TPB model (Table 3).
health training (X2(3) ¼16.853, p < .01) was detected. There was no significant relationship between
Of those coaches who had previous training, 77% Intention and the number of athletes (one vs. more
strongly agreed or agreed and 0% indicated neutrality than one) that a coach spoke with about mental
or disagreement towards the Intention measure. In health in the last season (p > .05)
comparison, in those coaches who had no previous The three TPB constructs were associated with expe-
training, 61% strongly agreed or agreed, and 22% riential factors. There was a significant difference in
expressed neutrality or disagreement. A reported 68% Attitude, PBC, and Social Norm scores between
of the total sample of coaches reported that they had a coaches with and without previous training as well as
conversation about mental health with their athletes those with and without personal experiences. Coaches
last season. Of those who reported speaking with ath- with previous mental health had significantly lower
letes, 42% reported speaking with only one athlete and Attitude [t(125) ¼ –1.981, p ¼ .05], PBC [t(125) ¼ –3.602,
58% reported speaking with more than one athlete. p < .001] and Social Norm [t(125) ¼ –2.579, p < .05]
There were no significant associations found between scores as compared to those with no training.
demographic or experiential factors and Behavior Similarly, coaches with personal experience had signif-
(p > .05). Similarly, there was no significant associa- icantly lower Attitude [t(133) ¼ –2.549, p < .05], PBC
tions between either demographic or experiential fac- [t(133) ¼ –2.494, p < .05] and Social Norm
[t(133) ¼ –2.501, p < .05] scores than those coaches with-
tors and if coaches spoke to one or more than one
out personal experience. There was also a significant
athlete about mental health (p > .05).
difference between age groups in Social Norm scores
The linear regression model with predictors Attitude,
with those 40 years and under reporting significantly
PBC and Social Norms explained a significant amount
lower scores than those 41 years and over [t(132) ¼ –
(42.7%) of variance in Intention. The model identified
2.250, p < .05]. Lastly, those coaches who reported
that all three constructs had a significant positive rela-
talking to their athletes had significantly lower PBC
tionship with Intention. In other words, an increase in
scores [t(128) ¼ 3.642, p < .001] and Social Norm
construct scores corresponds with an increase in
Intention score. This means that as Attitude, PBC and

Table 2. Summary of theory of planned behavior regression models.

Linear regression with intention

t p b F df p Adjusted R2

34.52 3 <.001 .427


Attitude 2.170 <.05 0.166
PBC 4.959 <.001 0.380
Social Norms 4.568 <.001 0.317
Logistic Regression with Behavior

B SE B Sig Exp(B) 95% CI for Exp(B)

PBC*Intention –.198 .075 <.01 0.820 0.708–0.951


Constant 1.047 .208 <.001 2.850
*denotes interaction.
PBC: perceived behavioral control.
514 International Journal of Sports Science & Coaching 16(3)

Figure 2. Interaction plots illustrating the effect of pbc on the relationship between talked and role.
It was created using RStudio41

Table 3. Correlation between theory of planned behavior


constructs. whether a coach spoke to their athletes about mental
health. The objective was two-fold, firstly, identify if
1 2 3 Attitudes, PBC, and Social Norms could predict
Attitude 1.000 Intention and secondly, to measure the interaction
PBC 0.502* 1.00 effect of Intention and PBC on Behavior.
Social Norms 0.303* 0.297* 1.00 The current study found that 70% of coaches agreed
or strongly agreed that it was within their role to sup-
*Significant correlation (p < 0.01).
PBC: perceived behavioral control.
port athlete mental health. Approximately the same
percentage (68%) of coaches reported that they had
talked to their athletes about mental health in the last
[t(128) ¼ 3.839, p < .001] scores as compared to those season. Results suggested that the model structure pro-
who did not report talking (Table 4). posed by the TPB fits the sampled data. Attitude, PBC
and Social Norm scores explained approximately 43%
Discussion of the variance in coach Intention. The significant pos-
itive relationship indicated that as coach construct
Recently published statements have called on coaches scores increased (became less supportive) coaches
to be involved in and support athlete mental were less likely to agree that it is within their role to
health.14,16 While there may be many ways for a be involved in athlete mental health. Further regression
coach to do so, research has highlighted the benefits analysis revealed a significant interaction effect of PBC
of conversation and open discussion around mental scores and Intention on Behavior (Figure 2). This sig-
health.14–16 Mazzer and Rickwood21 hypothesized nificant negative relationship suggested that as the
that rewards/beneficial outcomes, perceived capabili- interaction score increased, coaches were less likely to
ties, role requirements and expectations of others may report having performed the Behavior. Such results
help explain coach behaviors. Given the highlighted follow the TPB model, which states that there is a sig-
importance of conversation, the purpose of this study nificant effect of the interaction between perceived con-
was to apply the TPB to measure factors that influence trol over the behavior and intention to perform the
Murphy and Sullivan 515

Table 4. Average attitude, perceived behavioral control and social norm scores by group.

Attitude PBC Social Norms

Demographic variable n x ðSDÞ n x ðSDÞ n x ðSDÞ

Gender
Male 83 2.13 (0.72) 83 2.35 (7.50) 83 2.60 (0.93)
Female 50 1.90 (0.60) 50 2.35 (0.75) 50 2.40 (0.84)
Talked
Yes 93 1.96 (0.67) 93 2.15 (0.66)* 93 2.31 (0.79)*
No 37 2.14 (0.69) 37 2.64 (0.77) 37 2.93 (0.95)
Previous mental health training
Yes 63 1.91 (0.60) 63 2.11 (0.65) 63 2.31 (0.87)
No 64 2.15 (0.72) 64 2.56 (0.77) 64 2.73 (0.89)
Age
40 71 2.07 (0.69) 71 2.28 (0.75) 71 2.36 (0.88)
 41 63 2.03 (0.69) 63 2.35 (0.73) 63 2.71 (0.91)
Personal experience
Yes 101 1.95 (0.66)§ 101 2.21 (0.70)§ 101 2.40 (0.87)§
No 34 2.29 (0.72) 34 2.56 (0.74) 34 2.84 (0.93)
*Significantly different than “No” (p < 0.05);
,
Significantly different than “No” Training (p ¼ 0.05);
,
Significantly different than “No” Training (p < 0.05);

Significantly different than “  41” (p < 0.05);
§
Significantly difference than “No” experience (p < 0.05);
Comp: Competitive; n: number of participants; PBC: perceived behavioral control; SD: standard deviation; 
x : mean.

behavior on the actual performance of the behavior in training may be more likely to agree that their role
question. includes being involved in athlete mental health. It is
The current results were inconsistent with earlier possible that the benefits of mental health training,46
research in terms of the relationship between demo- explained further below, impacts role perception of
graphic factors, Intention and Behavior. Earlier coaches. However, it is also possible that those coaches
research in high school coaches, has reported that who seek out training already have these role
older coaches were less likely to help than younger perceptions.
coaches.44 The current study found no difference in In addition to the association between previous
Behavior between age groups; this lack of difference mental health training and Intention, training also
may be caused by measurement; age was measured by had a positive association with Attitude, PBC and
group, and not precise age. Additionally, previous lit- Social Norm scores. Various forms of mental health
erature with college-level coaches has found gender to education have been shown to reduce stigma levels,
be a significant predictor of intention to support, with improve knowledge and increase confidence to help
females showing more intention to perform these someone with a mental disorder.46 The difference in
behaviors.45 In the current sample, there was no asso- construct scores between trained and untrained coaches
ciation between gender and Intention, however, results may indicate that the training sessions received by these
suggest that with either a larger sample size, or a more coaches are effective. The association between training
balanced ratio of males to females, a difference may and PBC, specifically, supports earlier claims of an
have occurred. For instance, 18% of females reported association between training and preparedness to talk
either somewhat agreeing, being neutral or disagreeing and have open dialogue with athletes in intercollegiate
that their role as a coach was to be involved in athlete coaches.22 A similar benefit was also found for person-
mental health. In comparison, 49% of male coaches al experience; coaches with experience had better
reported the same opinion. Although no demographic Attitude, PBC, and Social Norm scores as compared
factors had an association with Intention, an associa- to coaches with no experience. Literature suggests
tion between Intention and previous mental health that individuals with experience often show higher
training was found. This suggests that as compared mental health literacy, lower stigmatization and lower
to their untrained colleagues, coaches with previous social distancing desires as compared to those with no
516 International Journal of Sports Science & Coaching 16(3)

experience.47–51 In intercollegiate coaches specifically, the TPB constructs, or the behavior itself. Third, the
it has been reported that personal experiences are data from this study does not give any detail on the
used to better understand and support athletes.22 depth of conversation or the exact topic of the conver-
Lastly, younger coaches (40 years) had significantly sation had between athletes and their coaches. As such,
lower Social Norm scores as compared to older ( the results can only be used to suggest factors that are
41 years) coaches. Older coaches may believe their col- associated with the act of having the conversation, not
leagues and athletes are less supportive of conversa- the type of conversation that is had. Future research
tions about mental health as compared to younger should investigate this further. Lastly, the Cronbach
coaches. This finding supports earlier research report- alpha scores for the individual survey sections was
ing a significant negative association between age and not optimal. Future studies could add more questions
mental health literacy in general populations52 and to the Social Norms section by adding perceptions of
intercollegiate coaches.53 The lack of differences in family, friends) and asking how important the opinions
construct scores between genders is somewhat contra- of these individuals are to the coach. Despite these
dictory to earlier research. Earlier research in intercol- limitations, the current study offers novel information
legiate athletes found female coaches to have to the area.
significantly greater mental health literacy as compared Although a promising amount of coaches reported
to males. As such, it was somewhat unexpected to find speaking with their athletes about mental health, there
no significant difference between genders. is still room for improvement. As this study helps to
Although demographic and experiential variables confirm the benefits of mental health training, such
did not relate to Behavior, data did reveal that individ- education should continue to be promoted to all mem-
ual construct scores may. Results indicated that those bers of the athletic community at all levels of sport.
coaches who talked to their athletes had significantly Results indicate that such training helps to improve
lower PBC and Social Norm scores as compared to attitudes, perceived capabilities, and social norms. In
those coaches who did not talk. Such results support addition, the significant association between training
the hypothesis by Mazzer and Rickwood21 that helpful
status and Role in combination with the association
behaviors are impacted by perceived capabilities and
between Role and Talked, implies their may be a
whether the behavior is expected or required within
direct association between training and Talking. In
the work role. However, contrary to this hypothesis,
general, attitudes were quite positive; as such, while
Attitude scores were not significantly different between
training should still promote such beliefs, it may be
coaches who talked and coaches who did not talk. The
beneficial to focus heavily on capabilities and perceived
association between attitude and behaviors docu-
social norms within sport. It is important to continue
mented in earlier research is inconsistent; personal atti-
endorsing positive attitudes as the sampled population
tudes towards help-seeking were shown to have a
may be biased towards such positive beliefs. Perceived
significant positive association with coach support in
college-level coaches45 but not in high-school capabilities and social norms appear to relate to wheth-
coaches.44 The difference between the current intercol- er a coach talks to their athletes or not; social norms,
legiate coach sample and that in the study by Kroshus also seem to vary most between individuals. By con-
and colleagues45 may be due to the types of attitudes tinuing to offer training to all individuals within the
that were measured. In the current study, attitudes athletic environment it may be possible to not only
towards the outcome of the behavior were measured improve skills but also to improve social norms. This,
as opposed to personal attitudes towards help-seeking in turn, may allow coaches to feel more prepared to
measured by Kroshus and colleagues.45 Further, in the help their athletes, and make the action of helping
current sample, the lack of significance may be due to feel more socially acceptable.
the overall low (positive) Attitude score. It is possible
that this sample had resoundingly positive attitudes Declaration of conflicting interests
towards speaking about mental health with athletes. The author(s) declared no potential conflicts of interest with
The conclusions drawn from the current study respect to the research, authorship, and/or publication of this
should be interpreted under the consideration of the article.
following limitations. Firstly, Attitudes, PBC, and
Social Norms of coaches reflected their current state, Funding
while the Behavior was retrospective. In addition, the The author(s) received no financial support for the research,
sample size was quite small and the ratio of males to authorship, and/or publication of this article.
females was imbalanced. Furthermore, all measures
were self-reported, as such, it is possible that social ORCID iD
desirability may have influenced responses for any of Jessica Murphy https://orcid.org/0000-0003-2732-8498
Murphy and Sullivan 517

References 16. Schinke RJ, Stambulova NB, Si G, et al. International



1. Arnett JJ, Zukauskien_ e R and Sugimura K. The new life society of sport psychology position stand: athletes’
stage of emerging adulthood at ages 18-29 years: impli- mental health, performance, and development. Int J
cations for mental health. Lancet Psychiatry 2014; 1: Sport Exer Psychol 2018; 16: 622–639.
569–576. 17. Berg BK and Warner S. Advancing college athlete devel-
2. National Institute of Mental Health (NIMH). Mental opment via social support. Journal of Issues in
illness, https://www.nimh.nih.gov/health/statistics/ Intercollegiate Athletics 2019; 12: 87–113.
18. Biggin IJR, Burns JH and Uphill M. An investigation of
mental-illness.shtml (2019, accessed 24 July 2019).
3. Jaworska N, De Somma E, Fonseka B, et al. Mental athletes’ and coaches’ perceptions of mental ill-health in
health services for students at postsecondary institutions: elite athletes. J Clin Sport Psychol 2017; 11: 126–147.
19. Way WC, Coker-Cranney AM and Watson IIJ. “So
a national survey. Can J Psychiatry 2016; 61: 766–775.
many mental health issues go unsaid”: implications for
4. Giovannetti S, Robertson JR, Colquhoun H, et al.
best practice guidelines from student-athletes’ perspec-
Mental health services for Canadian university student-
tives about service availability. J Clin Sport Psychol
athletes: an exploratory survey. J Clin Sport Psychol
2020; 14: 305–324.
2019; 13: 469–485.
20. Mazzer KR and Rickwood DJ. Mental health in sport:
5. Gulliver A, Griffiths KM and Christensen H. Barriers
coaches’ views of their role and efficacy in supporting
and facilitators to mental health help-seeking for young
young people’s mental health. Int J Health Promot
elite athletes: a qualitative study. BMC Psychiatry. Epub
Educ 2015; 53: 102–114.
ahead of print 2012. DOI: 10.1186/1471-244X-12-157.
21. Mazzer KR and Rickwood DJ. Teachers’ and coaches’
6. Van Slingerland KJ, Durand-Bush N and Rathwell S.
role perceptions for supporting young people’s mental
Levels and prevalence of mental health functioning in
health: multiple group path analyses. Aust J Psychol
Canadian university student-athletes. Can J High Educ
2015; 67: 10–19.
2018; 48: 149–168.
22. Rahill C. College coaches’ experiences, knowledge and
7. Sullivan P, Blacker M, Murphy J, et al. Levels of psycho-
attitudes to support student-athlete mental health.
logical distress of Canadian university student-athletes.
University of Vermonth ScholarWorks @ UVM
Can J High Educ 2019; 49: 47–59. Graduate College Dissertations and Theses, http://sch
8. Castaldelli-Maia JM, de Mello e Gallinaro JG, Falcao
olarworks.uvm.edu/graddis/1241 (2020).
RS, et al. Mental health symptoms and disorders in 23. Mazzer KR and Rickwood DJ. Community-based
elite athletes: a systematic review on cultural influencers roles promoting youth mental health: comparing the
and barriers to athletes seeking treatment. Br J Sports roles of teachers and coaches in promotion, prevention
Med 2019; 53: 707–721. and early intervention. Int J Mental Health Promot 2013;
9. Moreland JJ, Coxe KA and Yang J. Collegiate athletes’ 15: 29–42.
mental health services utilization: a systematic review of 24. Ajzen I. The theory of planned behavior. Org Behav Hum
conceptualizations, operationalizations, facilitators, and Decis Process 1991; 50: 179–211.
barriers. J Sport Health Sci 2018; 7: 58–69. 25. Godin G and Kok G. The theory of planned behavior: a
10. Ryan H, Gaston Gayles J and Bell L. Student-athletes review of its applications to health-related behaviors. The
and mental health experiences. New Directions for science of health promotion: behavior. Am J Health
Student Services 2018; 2018: 67–79. Promot 1996; 11: 87–98.
11. National Collegiate Athletic Association (NCAA). 26. Tye J. Predictors of faculty intentions to refer students
NCAA GOALS study of the student-athlete experience: with mental health concerns to mental health professio-
initial summary of findings, https://www-ncaa-org.proxy. nals. Culminating Projects in Higher Education
library.brocku.ca/sites/default/files/GO (2016, accessed 6 Administration 2016; 9: 1–170.
October 2020). 27. Schwartz LS. Faculty role in responding to the acutely
12. Wolanin A, Hong E, Marks D, et al. Prevalence of clin- distressed college student. ProQuest Dissertations
ically elevated depressive symptoms in college athletes Publishing, The George Washington University, USA,
and difference by gender and sport. Br J Sports Med 2010.
2016; 50: 167–171. 28. Mazzer KR and Rickwood DJ. Community gatekeepers’
13. Sebbens J, Hassmen P, Crisp D, et al. Mentla health in advice to young people to seek help from mental health
sport (MHS): improving the early intervention, knowl- professionals: youth workers and sport coaches. Int J
edge and confidence of elite sport staff. Front Psychol Ment Health Promot 2009; 11: 13–23.
2016; 7: 1–9. 29. Kroshus E, Baugh CM, Daneshvar DH, et al.
14. Henriksen K, Schinke R, Moesch K, et al. Consensus state- Understanding concussion reporting using a model
ment on improving the mental health of high performance based on the theory of planned behavior. J Adolesc
athletes. Int J Sport Exer Psychol 2020; 18: 553–560. Health 2014; 54: 269–274.
15. National Collegiate Athletic Association N. Supporting 30. Kroshus E, Baugh CM, Daneshvar DH, et al.
student-athlete mental wellness, http://www.ncaa.org/ Concussion reporting intention: a valuable metric for
sport-science-institute/supporting-student-athlete-menta predicting reporting behavior and evaluating concussion
l-wellness (accessed 19 June 2019). education. Clin J Sport Med 2015; 25: 243–247.
518 International Journal of Sports Science & Coaching 16(3)

31. Carpenter S, Lininger M and Craig D. Intrapersonal fac- 43. Kim H-Y. Statistical notes for clinical researchers: assess-
tors affecting concussion reporting behaviors according ing normal distribution (2) using skewness and kurtosis.
to the theory of planned behavior in high school football Restor Dent Endod 2013; 38: 52–54.
players. Int J Sports Phys Ther 2020; 15: 374–379. 44. Kroshus E, Chrisman SD, Coppel D, et al. Coach sup-
32. Rigby J, Vela L and Housman J. Understanding athletic port of high school student-athletes struggling with anx-
trainers’ beliefs towards a multifaceted sport-related con- iety or depression. J Clin Sport Psychol 2019; 13:
cussion approach: application of the theory of planned 390–404.
behavior. J Athl Train 2013; 48: 636–644. 45. Kroshus E, Wagner J, Wyrick DL, et al. Pre-post evalu-
33. QualtricsXM. QualtricsXM, https://www.qualtrics.com/, ation of the ‘Supporting Student-Athlete Mental
2020. Wellness’ module for college coaches. J Clin Sport
34. Corrigan PW, Morris SB, Michaels PJ, et al. Challenging Psychol 2019; 13: 668–685.
the public stigma of mental illness: a meta-analysis of 46. Breslin G, Shannon S, Haughey T, et al. A systematic
outcome studies. Psychiatr Serv 2012; 63: 963–973. review of interventions to increase awareness of mental
35. Griffiths KM, Carron-Arthur B, Parsons A, et al. health and well-being in athletes, coaches and officials.
Effectiveness of programs for reducing the stigma associ- Syst Rev 2017; 6: 1–15.
ated with mental disorders. A meta-analysis of randomized 47. Farrer L, Leach L, Griffiths KM, et al. Age differences in
controlled trials. World Psychiatry 2014; 13: 161–175. mental health literacy. BMC Public Health 2008; 8: 125.
36. Bradbury A. Mental health stigma: the impact of age and 48. Furnham A, Annis J and Cleridou K. Gender differences
gender on attitudes. Community Ment Health J 2020; 56: in the mental health literacy of young people. Int J
933–938. Adolesc Med Health 2014; 26: 283–292.
37. Ajzen I and Fishbein M. The influence of attitudes on 49. O’Connor M, Casey L and Clough B. Measuring mental
behavior. In: Albarracin D, Johnson BT and Zanna MP health literacy - a review of scale-based measures. J Ment
(eds) Handbok of attitudes an attitude change: basic prin- Health 2014; 23: 197–204.
ciples. Mahwah: Erlbaum, 2005, pp. 117–138. 50. Svensson B and Hansson L. How mental health literacy
38. Connor M. The theory of planned behavior. In: and experience of mental illness relate to stigmatizing
Tenenbaum G and Eklund RC (eds) Handbook of sport attitudes and social distance towards people with depres-
psychology. Hoboken: John Wiley & Sons, Inc., 2020. sion or psychosis: a cross-sectional study. Nord J
39. Hankins M, French D and Horne R. Statistical guide- Psychiatry 2016; 70: 309–313.
lines for studies of the theory of reasoned action and the 51. White M and Casey L. Helping older adults to help them-
theory of planned behaviour. Psychol Health 2000; 15: selves: the role of mental health literacy in family mem-
151–161. bers. Aging Ment Health 2017; 21: 1129–1137.
40. IBM. SPSS Statistical Software Version 26. USA, https:// 52. Hadjimina E and Furnham A. Influence of age and
www.ibm.com/analytics/spss-statistics-software?lnk= gender on mental health literacy of anxiety disorders.
STW_US_STESCH&lnk2=trial_SPSS&pexp=def& Psychiatry Res 2017; 25: 8–13.
psrc=none&mhsrc=ibmsearch_a&mhq=spss, 2020. 53. Sullivan P, Murphy J and Blacker M. The level of mental
41. RStudio Team. RStudio: Integrated Development for R. health literacy among athletic staff in intercollegiate
Boston, MA: R Studio PBC, 2020. sport. J Clin Sport Psychol 2019; 13: 440–450.
42. Tavakol M and Dennick R. Making sense of Cronbach’s
alpha. Int J Med Educ 2011; 2: 53–55.

You might also like