Exploring nonverbal behavior in elite handball players: development of the Handball Post-Shot Behavior Coding Scheme (H-PSB-CS).
Moesch, Karin ; Kentta, Goran ; Mattsson, C. Mikael 等
Matches in team sport competitions normally last at least an hour
and involve many situations in which athletes attempt to score. These
attempts can result in either success (i.e. score) or failure (i.e.
shots that do not result in a goal). Based on the outcome of each
specific situation, players usually show different forms of behaviors as
a reaction to their attempts. In the context of shooting to the
goal--which can be considered the most critical moment in the game--such
behaviors have been labeled post-shot behaviors (Moll, Jordet, &
Pepping, 2010). These behaviors can be verbal (e.g., shouting or giving
feedback to a teammate) and/or nonverbal. The latter include cues from
different channels, such as facial expressions, posture, gestures, or
touch (see Riggio & Riggio, 2012), which in the presence of others
can function as nonverbal communication messages (Richmond, McCroskey,
& Hickson III, 2012). Burgoon, Buller, and Woodall (1996) argue that
people tend to rely more heavily on nonverbal, compared to verbal,
communication in times of stress. Competitive events are generally
considered to entail many stressors for athletes (e.g., Mellalieu, Neil,
Hanton, & Fletcher, 2009). Moreover, verbal communication is not
always possible during an on-going match (e.g., due to size of the
court, or noise level in the arena); thus the nonverbal part of the
communication becomes even more salient. The focus of the present study
will be on nonverbal behaviors in the post-shot period during team sport
matches.
Several functions of nonverbal behavior exist (Baesler &
Burgoon, 1987), some of which are highly important in the team sport
context. For example, expressing emotions is one area of nonverbal
behavior (Baesler & Burgoon, 1987), and Riggio and Riggio (2012)
even stated that emotions are primarily communicated through nonverbal
cues. Some of the channels mentioned above, for example gestures and
facial expressions, are supposed to be specifically related to the
display of emotions (Bull & Doody John, 2013; Kappas, Krumhuber,
& Kiister, 2013). There is evidence suggesting that some emotions
are predominantly communicated through specific channels: social status
emotions through the body channel, survival emotions through facial
expressions, and intimate emotions through touch (App, McIntosh, Reed,
& Hertenstein, 2011). Expressing emotions is considered to fulfill
an important function by quickly and nonverbally communicating social
information in complex societies (Shariff & Tracy, 2011). In the
sport context, the nonverbal behavior seen in the post-shot period may
display the spontaneous expressions of joy of the player who scored,
which is communicated to teammates through specific behaviors. However,
at times, emotional expressions do not necessarily reflect spontaneous
emotional states. As such, socially learned rules that dictate the
management of affect display in social settings exist (cf. display
rules, see Ekman & Friesen, 1969) and may at times modulate the
spontaneous emotional expressions. In a team sport context, for example,
a disappointed player might suppress her spontaneous emotional reaction
(e.g., tramping into the court in frustration), as she knows that such
behavior would not be accepted in this specific situation. Merging these
two perspectives, a study by Tracy and Matsumoto (2008), comparing
behavioral expressions of blind, congenitally blind and sighted winning
judoka athletes, found that all three groups showed similar spontaneous
behavioral expressions associated with shame and pride that are thus
likely to be innate; however, it also emerged that the shame display is
intentionally inhibited by sighted athletes in accordance with learned
cultural norms.
Linked to the idea of (learned) display rules in the context of
emotional expressions, it can also be assumed that nonverbal behavior
might simply be the result of a learning process, without a player
experiencing specific emotions. As suggested by Bandura (1977),
behaviors can be learned through modeling others. Specifically, a player
might observe more experienced players showing certain behaviors after
being successful, and copy these behaviors into their own behavior
repertoire. Following this argumentation, the shown behavior is
considered a ritual, but not necessarily the display of a specific
emotion. Another possible reason for the display of nonverbal behavior
could be that players show them for a very specific reason, such as
cheering their own teammates, daunting their opponents, or pleasing
supporters. In these cases, nonverbal behaviors are consciously
performed in order to attain a certain result. The functions of
nonverbal behavior can thus be considered to be diverse.
Subsequently, one can wonder what impact such nonverbal behaviors
can have in a team sport context. Based on empirically supported
theoretical foundations, displaying nonverbal behavior is hypothesized
to serve several functions. First, the athlete is assumed to be affected
by his or her own emotional expression through internal feedback loops
(e.g., Price, Peterson, & Harmon-Jones, 2012). Translated into the
sport context, this argumentation suggests that a player celebrating a
successful action (e.g., through gestures implying joy) will cultivate
or intensify this emotional state through an internal feedback loop.
Second, teammates (the "in-group", see Sherif, 1966) are
influenced by a player's emotional expression through a mechanism
called emotional contagion (e.g., Hatfield, Cacioppo, & Rapson,
1994). Within the sport context, showing positive emotions after a
successful action is supposed to influence the emotion of teammates, and
so far some evidence exists that supports this assumption (Moll et al.,
2010; Totterdell, 2000). Thirdly, players form perceptions of opponents
(the "out-group", see Sherif, 1966) based on nonverbal
behavior, which influences their judgment about the opponent's
ability, and finally their own affective response (Warr & Knapper,
1968). Several studies support the importance of different nonverbal
cues on the formation of impression in a sport context (Furley &
Dicks, 2012; Furley, Dicks, & Memmert, 2012; Greenlees, Bradley,
Holder, & Thelwell, 2005; Greenlees, Buscombe, Thelwell, Holder,
& Rimmel, 2005).
Besides the aforementioned theoretical reflections, there is
considerable knowledge from qualitative studies indicating that
nonverbal behavior play an important role in team sports. In one such
study, handball players perceived that positive nonverbal behavior
intensified the opponent's feeling of defeat (Ronglan, 2007).
Moreover, the handball players also considered cheering for each other
or expressing collective joy during a competitive event as being crucial
for collective efficacy and success (Ronglan, 2007). In line, Fransen et
al. (2012) showed that supportive communication, such as coming together
enthusiastically after points or cheering by field and bench players
after points, was the most predictive factor for positive collective
efficacy beliefs in volleyball. Meanwhile, a negative emotional
reaction, such as expressing discouraging body language, was the most
predictive factor for negative collective efficacy beliefs. Moesch and
Apitzsch (2012) revealed that handball coaches considered showing
positive reactions within the team as a useful strategy to enhance
positive psychological momentum in a team, while negative body language
is considered a trigger of negative psychological momentum. A similar
study within psychological momentum and soccer showed that negative body
language of the opponent functions as a trigger for positive momentum of
the own team, and, correspondingly, negative body language of the own
team was considered to be a result of being in a negative psychological
momentum (M. Jones & Harwood, 2008). These results show that
nonverbal behavior have been connected to different team-related
constructs.
Even though a broad array of theoretical explanations exist, as
well as some qualitative studies highlighting the importance of
nonverbal behaviors in a team sport setting, only a few quantitative
studies have been made that explore nonverbal behaviors in real-life
setting within sports. Matsumoto and Willingham (2006) focused on facial
expressions of Olympic medal winners, using the well-known Facial Affect
Coding System (FACS; Ekman & Rosenberg, 2005). Additionally, three
studies exist on touching behavior in the post-competitive period
(Anderton & Heckel, 1985; Heckel, 1993; Heckel, Allen, &
Blackmon, 1986). Overall, the focus of these studies was on nonverbal
behaviors resulting from a competitive event; few researchers have taken
the challenge to investigate nonverbal behaviors that include various
components such as touch, gesture, and facial expressions during an
on-going game (Bornstein & Goldschmidt, 2008; Kneidinger, Maple,
& Tross, 2001; Kraus, Huang, & Keltner, 2010; Moll et al.,
2010).
Several reasons can be assumed to explain the rather sparse
research tradition within nonverbal behavior in sports, especially
during games. To investigate nonverbal behavior, observational studies
are considered to be the most favorable design (Hertenstein, Verkamp,
Kerestes, & Holmes, 2006), but such a design also involves some
challenges. First, it is generally acknowledged that observational
studies require a lot of resources (Furr & Funder, 2007; Gee &
Sullivan, 2006). Second, an arguably even bigger barrier is that
suitable coding schemes that enable the reliable assessment of nonverbal
behaviors in a sporting context are sparse. One of the few examples of
such a coding scheme stems from Moll et al. (2010) who developed a
coding scheme based on research from Tracy and Robins (2007a) that
included behaviors associated with pride and shame. This provided the
first possibility to code nonverbal behavior associated with different
channels (body and face) in the sport context; however, their study
included a rather static game situation (penalty shootouts in soccer)
that differs markedly from the flow state found during normal match time
in any team sport. As such, it has to be questioned if their suggestions
of nonverbal behaviors in the post-shot period can be measured reliably
during normal match time. Coding schemes developed in studies by
Kneidinger et al. (2001) and Kraus et al. (2010) offer interesting
information about the study of one possible channel, namely touch, in
team sport contexts. Finally, a study by Bornstein and Goldschmidt
(2008) focused among others on the amount of touch the scoring player
received, but without further specifying the form of the touch
displayed. Alternatively, existing material on coding of emotional
expressions in general psychology (Cohn & Ekman, 2005; Dael,
Mortillaro, & Scherer, 2012a, 2012b; Ekman & Friesen, 1972;
Ekman & Rosenberg, 2005; Friesen, Ekman, & Wallbott, 1979) can
provide a useful and important base, but is not easily applicable to the
world of sports. For instance, Kneidinger et al. (2001) suggest that
display rules are loosened in the sport context, which leads to a
different array of emotional expressions than are exhibited in public
settings. Within the sport context it is further suggested that
emotional expressions shown during competition differ markedly from one
sport to another (Cerin, Szabo, Hunt, & Williams, 2000). This can be
explained by the format of the game, which can include varying
possibilities for players of different team sports to display nonverbal
behavior (e.g., a soccer goal is followed by a break of the game,
whereas a goal in handball does not interrupt the flow of the game).
Additionally, a goal in a sport where few goals happen (e.g., soccer or
ice hockey) would arguably trigger a stronger reaction than in a sport
where many goals usually happen (e.g., handball or basketball). As a
result of these reflections, it seems essential to study nonverbal
behaviors in the post-shot period within one single sport, and develop
specific coding schemes in accordance to this sports' unique
display rules. Another aspect that has to be considered when developing
coding schemes are gender factors; it is stated that men and women
display emotions with differing intensity (see e.g., Brody & Hall,
2008), and that there are differences between males and females when it
comes to nonverbal behavior in general, and, in specific, facial and
gestural expressiveness (Hall & Gunnery, 2013; Richmond et al.,
2012). This forces researchers interested in the study of nonverbal
behavior to focus on a specific gender in order to attain high face
validity.
The present study aims to develop a coding scheme that adequately
measures female handball players' nonverbal behaviors in the
post-shot period. Handball has been chosen as it is a sport where many
goals happen in each game (and thus consequently many post-shot periods)
and is played on a court with a size that is easily observable for
coding. We claim that information about the development of a coding
scheme will be of particular interest for research colleagues who want
to shift toward more behavior-focused research within the area of
nonverbal behavior in team sports. Once a solid coding scheme is
developed, more applied research questions, such as the description of
nonverbal behaviors in general and for winning or losing teams, or the
impact of nonverbal behaviors for subsequent performance, can be
studied. Moreover, a sound and reliable coding scheme will likewise be
of interest for sport psychology practitioners who want to monitor and
optimize nonverbal behavior in their teams.
Method
Design
A naturalistic design with systematic observation was chosen for
the present study. Such a design enables the gathering of rich data of
the phenomena in question within a real-world setting with high
ecological validity (Hertenstein et al., 2006; Stangor, 2007). The
target behavior was videotaped from the public area, thus adopting a
nonparticipant observer perspective, which is generally acknowledged to
be unobtrusive and minimizes reactivity in the observed persons (Cooper,
Heron, & Heward, 2007; Miltenberger & Weil, 2013; Stangor,
2007). This design also offers the opportunity to gather information
about nonverbal behaviors that are usually considered to be out of our
awareness (Riggio & Riggio, 2012) and thus difficult to gather with
self-report methods.
Data
Matches. Data were gathered at matches of the highest female
handball league in Sweden during the 2011/2012 season. As it is
generally acknowledged that emotions are more frequent when high value
goals are at stake (Lazarus, 1999), it was hypothesized that during such
matches a high amount of nonverbal behaviors would be displayed.
Therefore, data material were gathered from matches that were judged in
advance by a handball expert (third author) to be matches with a high
stake, defined as: matches played between teams from the same
city/region ("derbies"), matches played between teams that
were close to each other in the ranking, or play-off matches. Finally,
matches were selected based on arenas with good light conditions for
filming and within reasonable travelling distance for the three
cameramen involved. Eighteen matches (eight from the official league, 10
from the play-offs) were filmed exclusively for the purpose of this
study, resulting in a total of 997 minutes of filmed playing time.
Coding situation. The coding situation starts when a player
executes a shot with the intention to score and ends when she has
returned to her defense position. From the overall amount of coding
situations from the 18 matches, 114 coding situations were missing due
to different reasons (e.g., technical problems with the camera). These
114 coding situations stem from seven matches, ranging from 1 to 26
missing coding situations per match (M = 16.29, SD = 12.97). In total,
1,416 units of analysis were filmed, ranging from 45 to 98 per match (M=
78.47, SD = 13.58). From that number, 177 situations had to be deleted
due to problems with filming ([much less than] = 80, e.g., a teammate
blocked the view to the player in focus) or because the coding situation
was interrupted due to game related issues (n = 97, e.g., due to a
player getting injured, or a team taking a timeout directly after an
attack). In total, 1,239 situations are included in the analyses.
Materials
Furr and Funder (2007) recommend basing a coding scheme on both
practical and theoretical considerations. Therefore, existing literature
was reviewed, and four references were chosen to be most adequate for
the development of a coding scheme for the present study (Kneidinger et
al., 2001; Kraus et al., 2010; Moll et al., 2010; Tracy & Robins,
2007a). After that, videos of four matches from an international
handball tournament, especially filmed for the purpose of the present
study, were analyzed. Watching video films of the target population
helped the researchers to get familiarized with the observed behavior,
which is an important step in observer training (see e.g., Brewer &
Jones, 2002). During that process, a first version of the coding scheme
was elaborated.
Based on the recommendation of Brewer and Jones (2002), an expert
panel was organized in order to check the face validity of the coding
scheme. One female coach (having extensive experience as a player in the
highest leagues in Sweden and Norway and in the Swedish national team,
and working as a coach for several years) and three female players (all
playing in the highest league in addition to being members of the youth
([much less than] = 2) and the senior (n = 1) national team) were
invited to discuss different forms of post-shot behaviors. A nominal
group technique was chosen to guide the meeting (see e.g., Cantrill,
Sibbald, & Buetow, 1996; J. Jones & Hunter, 1995; Van de Ven
& Delbecq, 1972). The participants proposed five categories for
positive reactions (positive body language/gestures, focus on next task,
positive verbal communication, physical communication/touch, facial
expressions), and seven categories for negative reactions (negative body
language, physical expression of frustration, verbal outburst of
frustration, avoidance of communication, focus on wrong tasks,
displaying irritation towards teammates, rolling one's eyes).
However, as Miltenberger and Weil (2013) pointed out, behaviors should
not include category labels or internal states or characteristics.
Rather, it is recommended to define the target behavior in objective,
clear and complete terms including active verbs that describe the
person's observable behavior (see e.g., Kazdin, 2010; Miltenberger
& Weil, 2013). Doing so minimizes human error in the coding process,
which is considered to be the biggest threat to accuracy and reliability
in behavior analyses (Cooper et al., 2007). Therefore, only those
categories emerging from the expert panel including overt behaviors were
selected (positive body language/gestures, physical communication/touch,
negative body language, physical expression of frustration, and
displaying irritation towards teammates). Those behaviors, as well as
behaviors from the first version of the coding scheme, formed the basis
of the preliminary coding scheme. It included the three categories
positive post-shot behavior (PSB-P with 14 behaviors) that consisted of
gestures, facial and bodily expressions associated with positive
experiences (see table 1 a), negative post-shot behavior (PSB-N with 5
behaviors) that included gestures and bodily expressions associated with
negative experiences (see table 1 b), and post-shot behavior touch
(PSB-T with 8 behaviors) that consisted of different types of touch (see
table 1 c). Clear coding rules were formulated for the corresponding
behaviors to facilitate the coding process (see table 1 a-c). All data
coding was done with Sideline XPS Video Analyzer (version 12.5, Sideline
Sports, Iceland), a software that helped save the coding decisions in a
timely way and permitted coding situations to be replayed as many times
as necessary for complete coding.
Procedure
Coding consisted of registering all nonverbal behaviors of the
target player (i.e. the player who shot to the goal) during the coding
situation. It could include none of the behaviors, up to several
behaviors per category, and it was also possible to code the same
behavior several times during the same coding situation (e.g., when a
player raises a fist after scoring, runs back, and raises a fist again
when approaching the defense position). No behaviors from other players
than the target player were registered.
One important issue when conducting observational studies is
trustworthiness. Accuracy of data refers to the degree that an observed
value corresponds to the true value (Cooper et al., 2007), but is a
measure that is not possible to gather in observational studies as the
true value is unknown (Miltenberger & Weil, 2013). Reliability is
considered the next best quality indicator, and needs to be high in
order to optimize confidence in the data (Adler & Adler, 1994;
Cooper et al., 2007). Both observer training (Brewer & Jones, 2002;
Furr & Funder, 2007; Miltenberger & Weil, 2013) and inter- and
intra-observer agreement checks (Brewer & Jones, 2002; Cooper et
al., 2007; Miltenberger & Weil, 2013) are recommended in order to
enhance, and consequently prove, reliability of data.
The first author started by coding all 18 matches. Subsequently, a
meeting was held with the second author to discuss situations that were
difficult to code. During that meeting, 91 situations from 10 different
matches were thoroughly analyzed and discussed. As seven behaviors
(smile, laugh, erected body posture, slumped body posture, kicking in
the bench, kicking in the floor, hug) did not emerge in the data
material at all, it was decided to remove these behaviors from the
coding scheme. Some of these behaviors turned out not to be applicable
in handball (e.g., hugs) thus showing low content validity, some of the
behaviors did not occur during the time-frame of the coding situation
(e.g., kicking in the bench), and some behaviors were not possible to
code on video of moving subjects (e.g., slumped body posture, laugh). On
the basis of this meeting, the coding scheme was adapted, resulting in a
revised version of the coding scheme with 20 behaviors (11 PSB-P, 2
PSB-N, and 7 PSB-T). Moreover, the coding rules were further refined;
these coding rules formed the basis for the final analysis of the
complete data material by the first author. Subsequently, the revised
coding scheme was introduced to a research assistant (a Master student
in sport psychology with extensive experience in team sport) who was
blind to the research question (a naive observer, see Cooper et al.,
2007). During a meeting, the coding scheme was discussed. The research
assistant had to code 30 situations, and was thereafter provided with
feedback by the first author. All behaviors that did not emerge during
these 30 situations were presented from other matches in order to make
sure that the assistant researcher had seen all categories before
starting to code.
The assistant researcher coded nine randomly chosen matches,
including 620 coding situations. This corresponds to 50% of all data
material that was coded by two persons, and thus clearly lies above the
recommendation by Harrigan (2005) that typically 10-25% of the total
data material needs to be coded by several coders. During the analysis
of the two coding versions, it emerged that 38 situations had either an
agreement below 50%, or showed a different order of the same behaviors
during a coding situation. Therefore, these 38 situations were
re-analyzed by both coders together, a final decision was made on how to
code these coding situations (discrepancy discussion, see Yoder &
Symons, 2010), and they were subsequently eliminated from the agreement
analyses.
Additionally, the first author recoded two matches (randomly chosen
from the nine matches that were not coded by the research assistant) in
order to perform a test-retest reliability check. This retest was done
six weeks after completion of the final coding in order to ensure memory
lapse (Cooper et al., 2007). In total, 152 coding situations were
analyzed in the retest, corresponding to 12% of the complete data
material.
Analysis
A basic psychometric concern for observational studies is the
degree to which different observers, or one observer at different time
points, agree on the coding decision (Cone, 1977; Furr & Funder,
2007). Coefficients about inter-observer and intra-observer agreement
are then offered as reliability of the instrument being used (Boyce,
Carter, & Neboschick, 2000; Cooper et al., 2007; Mitchell, 1979).
The format of the present coding scheme, however, contains several
challenges that complicated these analyses. The revised version of the
coding scheme consisted of a total of 20 behaviors and it was neither
expected that all of those behaviors would be shown very often, nor that
every single coding situation would include one of the 20 behaviors
(e.g., a player may run back into defense without showing a nonverbal
behavior). Accordingly, a high amount of non-codable situations were
expected. Moreover, it was expected that some behaviors would be shown
several times under the same coding situation. Due to these arguments,
it was a) neither possible nor meaningful to count an agreement of the
non-occurrence of the different behaviors, and b) not possible to count
simple yes-no analyses, as the frequency of the behaviors was also
important. Due to these considerations, it was only possible to count
occurrence percentage agreement. This measure is the most frequently
used method in behavior analysis (Cooper et al., 2007; Mitchell, 1979;
Watkins & Pacheco, 2000; Yoder & Symons, 2010), even though it
has been criticized as not being an informative index of reliability as
it does not take into account a correction for chance (Baesler &
Burgoon, 1987; Rosenthal, 2005; Watkins & Pacheco, 2000). Focusing
on occurrence percentage agreement offered the possibility to count
inter- and intra-observer agreement from non-exhaustive agreement
matrices (Yoder & Symons, 2010), which is the case in the present
study.
Scored mean count-per-interval agreement was measured by counting
the average agreement over a complete coding situation where at least
one of the coders observed one behavior (Cooper et al., 2007). As this
procedure does not take into account if a coding situation consisted of
one or several behaviors, further analyses looked at the percentage
agreement that existed in all behaviors within a single category (PSB-P,
PSB-N, and PSB-T), and within all three categories together. Based on
the recommendation of Baesler and Burgoon (1987), behavior-specific
analyses were done in order to see if the agreement was consistent over
the different behaviors, or if any behavior was especially difficult to
code. The inter-observer agreement analyses include all coding
situations except the 38 situations that were discussed and agreed upon
by the two coders and due to that ended in a 100% agreement. Cooper et
al. (2007) state that a mean agreement of 75% is acceptable when
simultaneously coding multiple behaviors in complex environments; team
sport competitions are in fact highly complex environments. Moreover, as
handball players normally move quickly, and do so especially in the time
period the coding was done, an even lower agreement could result. Due to
that, the requirements above were adapted, in that for the behavior
specific analyses, the cut-off was set at an average agreement of 65% in
the two agreement tests (inter-observer and intra-observer). However, in
such a case, it needs to be investigated if the mistake was due to one
rater coding one behavior, and the other none, or if the mistake was due
to one rater coding one behavior, and the other rater coding a very
similar behavior (e.g., high versus low five). In the latter case, the
qualitative difference does not seem as serious, wherefore the agreement
level of 65% would be considered acceptable.
Results
Behavior-specific agreement
Within the category of PSB-P, the inter-observer agreement ranged
between 0% and 87%; meanwhile, the intra-observer agreement ranged
between 0% and 100% (Table 2). Four behaviors (one arm up, two arms up,
high tempo, fighting for the ball) showed a very low occurrence and/or a
very low agreement. It was therefore decided to delete these behaviors
from the coding scheme. Moreover, it emerged that the behavior airplane
resulted in a low inter-observer agreement. A closer look at the data
matrix revealed that eight of the nine misses in coding consisted of a
cross-coding with the category two fists up. Therefore, it was decided
to merge these two categories together for the final coding scheme. The
final inter- and intra-observer agreement when merging the two behaviors
was 88% and 84% respectively. The two behaviors one fist down and two
fists down resulted in a relatively low inter-observer agreement. The
data matrix revealed, however, that in both cases around 50% of the
miss-coding emerged from coding the respective behavior (one or two
fists) as high instead of low, which is considered to be of
qualitatively little difference. After the changes suggested above, the
category solely consisted of behaviors including gesture, wherefore it
was decided to re-name this category as post-shot behavior gesture
(PSB-G).
Among the behaviors for the category PSB-N, a low agreement was
found for both the behavior of low tempo (56 % resp. 20%) and displaying
expressions of frustration/irritation (40% resp. 100%; Table 2). Neither
behaviors met the criteria, and were therefore deleted from the coding
scheme. In the category PSB-T, the inter- and intra-observer agreement
ranged from 50% to 94%, and from 0% to 100%, respectively (Table 2). The
behaviors low ten and touch bottom resulted in an agreement below the
suggested criteria, leading to their deletion from the coding scheme.
The behaviors touch shoulders, and double touch also resulted in an
agreement that was questionable. However, the data matrix revealed that
the majority of the miss-codings in these categories stem from a
cross-coding with other, closely connected categories (e.g. a double
touch coded as a high five), which lead to the decision to leave these
categories in the coding scheme. At the end, the two categories PSB-G
with six behaviors (one fist down, two fists down, one fist up, two
fists up, thumbs up, clapping hands), and PSB-T with five behaviors (low
five, high five, high ten, touch shoulders, double touch), constituted
the final version of the coding scheme, labeled Handball Post-Shot
Behavior Coding Scheme (H-PSB-CS).
Category-specific agreement
The results of the analyses with the revised coding scheme revealed
an agreement in 402 of the 499 behaviors detected as PSB-P (81%), 11 of
the 23 in PSB-N (48%), and 404 of the 445 in PSB-T (91%). Counted over
all behaviors, an inter-observer agreement of 84% could be found, and
the scored mean-count-per interval agreement was 80%. The results of the
intra-observer analyses showed an agreement of 86% for the PSB-P, 67%
for the PSB-N, 95% for PSB-T, and an overall agreement of 89% for all
behaviors. The scored mean-count-per interval agreement was 86%.
After the changes in the coding scheme described above, the final
version of the coding scheme (H-PSB-CS) with six PSB-G and five PSB-T
was re-analyzed. The results revealed an inter-observer agreement of 84%
for PSB-G, 91% for PSB-T, and 87% for all behaviors together. The scored
mean-count-per interval agreement was 85%. For the intra-observer
agreement, the final results revealed an agreement of 86% for PSB-G, 95%
for PSB-T, and 90% for all behaviors together. The scored mean-count-per
interval agreement was 88% (Table 3). A table with the final categories
and behaviors is displayed in table 4.
Discussion
The quantitative investigation of nonverbal behavior during team
sports is a rather neglected area within sport psychology research. One
of the main reasons for this shortcoming is that few coding schemes are
available that reliably measure nonverbal behaviors exhibited by
athletes. Therefore, the aim of the present study was to develop a
coding scheme that adequately measures female handball players'
nonverbal behaviors in the post-shot period. Based on previous
literature, an expert panel, and thorough inter- and intra-agreement
analyses, the Handball Post-Shot Behavior Coding Scheme (H-PSB-CS),
including the categories post-shot behavior gesture (PSB-G) with six
behaviors (one fist down, two fists down, one fist up, two fists up,
thumbs up, clapping hands) and post-shot behavior touch (PSB-T) with
five behaviors (low five, high five, high ten, touch shoulders, double
touch), was developed and demonstrated to reliably capture nonverbal
behaviors in the post-shot period in female elite handball players.
During the process of developing a coding scheme, several
interesting points arose that are subject to discussion. At an early
stage in the development of the coding scheme, six behaviors in the two
categories PSB-P and PSB-N, namely smile, laugh, erected body posture,
slumped body posture, kicking in the bench and kicking in the floor,
emerged to be either unobservable in the post-shot period of handball
players, or impossible to code with acceptable reliability in this
complex context. The latter two behaviors (kicking in the bench and
kicking in the floor) emerged from the expert panel, but, nevertheless,
could not be detected during the coding situation. Therefore, they were
deleted from the coding scheme. Eventually, players who were banished to
the bench were involved in such behaviors, but it was not shown by
players during the coding situation. Another possible explanation is
that players of an elite level, such as those observed in this study,
are aware of the potential harmful effect of these types of behaviors on
team performance, and display rules, for example suppressing emotions
and looking affectless (Ekman & Friesen, 1969), come into effect and
prevent the display of true emotion. Detecting behaviors associated with
body posture (erected and slumped body posture), which has been
suggested to reveal emotions associated with pride, shame (Darwin, 1872;
Tracy & Robins, 2007a) and triumph (Matsumoto & Hwang, 2012),
proved to be impossible to observe when analyzing handball players in
motion and had to be deleted from the coding scheme. Similarly, Moll et
al. (2010) had to delete a behavior related to body posture (torso
pushed out) due to very low inter-observer agreement, and their results
further revealed that other behaviors including body posture generally
resulted in lower Kappa values compared with behaviors including
gestures. Furthermore, the behaviors smile and laugh had to be deleted
from the coding scheme as it was, due to the movement of the players,
not possible to detect these behaviors from video material with only one
camera. The study of Moll et al. (2010) detected both small and large
smiles, but these behaviors also resulted in lower Kappa values than
gestures. Their results further revealed that only few players showed a
large smile following a successful penalty. On one side, this is
surprising, as the smile has commonly been connected with happiness, joy
(Darwin, 1872; Ekman, 2003) and pride (Tracy & Robins, 2007a), and
been shown to occur as a result of athletic success (Matsumoto &
Willingham, 2006). However, recent research showed that a smile prior to
an athletic confrontation is a sign of lowered hostility and aggression,
and through that communicates reduced physical dominance (Kraus &
Chen, 2013). This raises the question whether smiling before or during a
competitive event is a functional behavior for athletes in sports where
opponents are directly faced. Research further revealed that the face is
one, but not the only channel by which emotions can be expressed (Tracy
& Robins, 2004a), and that some emotions are even better
communicated through other forms of nonverbal behavior, such as touch
(Hertenstein, Holmes, McCullough, & Keltner, 2009) or body posture
and movement (Tracy & Robins, 2004b). Thus, even though omitting
face codes in the present coding scheme seems an important shortcoming,
the aforementioned literature lends support to the notion that emotions
can be communicated through different channels. Moreover, capturing
facial expressions of athletes during on-going matches is still a big
challenge that is difficult to overcome in methodology.
Within the categories PSB-P and PSB-N, the final analyses with the
data material further lead to the insight that several behaviors are
shown too seldom in a handball specific context, or, alternatively, were
not possible to code with an acceptable level of reliability.
Specifically, fighting for the ball is a behavior that is infrequently
shown as a result of shooting to the goal; therefore, it was deleted
from the coding scheme. The two behaviors one and two arms up also
emerged very seldom; when arm movements are shown, they usually include
showing fists and were consequently coded as one or two fists up. Other
arm movements without fist(s) happened too infrequently to be included
in the coding scheme. It further emerged that is was too difficult to
code the two behaviors running back with high tempo (PSB-P) and running
back with low tempo (PSB-N); both behaviors are hard to differentiate
from a "normal" run back into defense, which highly depends on
game-related factors (e.g., when the opponent tries to make a fast
counter attack, the players are forced to run back faster than if the
goalkeeper first has to get the ball to continue playing). Even though
velocity of movement has been proposed to be associated with joy (high
movement dynamics) and sadness (low movement dynamics, see Wallbott,
1998), it was necessary to delete these behaviors as coding them did not
result in reliable data. Eventually, advancement in technology will
enable to measure such factors with specific GPS systems. The behavior
showing signs of frustration also resulted in a low agreement indicating
low reliability in coding, and therefore had to be deleted. Based on the
discussion above, it is apparent that the results of the present
investigation revealed that none of the behaviors originally suggested
in the PSB-N category could be reliably measured, or, alternatively,
were displayed only rarely during the coding situation.
Six behaviors, namely one fist down, two fists down, one fist up,
two fists up, thumbs up and clapping hands, emerged as typical and
reliably accessible behaviors in female elite handball. Making a fist
has been labeled as a sign of pride (Tracy & Robins, 2007a),
elevated joy (Wallbott, 1998), and triumph (Matsumoto & Hwang,
2012). In sport, showing fist(s) with one or two hands is a behavior
often shown by successful athletes. Likewise, raising arms has been
identified as a sign of elated joy (Wallbott, 1998), pride (Tracy &
Robins, 2007a), and triumph (Matsumoto & Hwang, 2012). For example,
Moll et al. (2010) found that celebrating a successful shot with both
arms was associated with winning the shootout. These authors
differentiated between codes for arm movements and hands in fists, while
in the present study arm movements could only be detected in combination
with fists. Thumbs up is a behavior that can be classified as an emblem,
which has been defined as a behavior with a direct verbal translation
that is used intentionally by the sender, and that is socially and
culturally learned (see e.g., Ekman & Friesen, 1969). In such, by
showing a thumb up to a teammate, a player can communicate a specific
meaning (e.g., communicating to a teammate that she did a good job by
passing the ball) when verbal communication is not possible. Such a
nonverbal behavior with a specific meaning can be an important positive
reinforcement that cultivates positive emotions in teammates. Moreover,
it has also been discussed that a thumbs up is a sign for triumph, thus
displaying a specific emotional state (Matsumoto & Hwang, 2012).
Clapping hands has been described as a sign of joy (Darwin, 1872) and
triumph (Matsumoto & Hwang, 2012); but it is also possible that a
player showed that behavior in order to communicate a specific message
to a teammate (e.g., when directing praise toward another player), or
used the behavior to pep herself up.
Within the category PSB-T, it emerged early in the process of
developing this coding scheme that hugs were not a behavior shown by
female elite handball players in the post-shot period, and, as such, was
deleted from the coding scheme. This is in contrast with other studies
within sport: personal hugs and team hugs are shown in the studies of
Kneidinger et al. (2001) and Kraus et al. (2010) in softball/baseball
and basketball teams respectively. The specific context of a
softball/baseball match differs significantly from a handball match,
allowing for different forms of celebratory behaviors and offering a
potential explanation for the differences found. However, it is rather
surprising that even in a basketball setting, being much more comparable
with handball, this behavior is shown. Eventually, the specific culture
in some sports reinforces (team) hugs, whereas in others it does not.
Alternatively, the different timeframe of data collection in the studies
(post-shot period versus complete match) may explain these results.
After the final analyses, it emerged that the behaviors low ten and
touch bottom were shown very rarely. Neither Kraus et al. (2010) nor
Kneidinger et al. (2001) coded the low ten, even though it was suggested
in the original coding scheme of the latter study, and while low fives,
high fives, and high tens appeared in both studies and also emerged from
the current results, it seemed that team sport athletes in general do
not show low tens during a game. The behavior touching bottom appeared
in softball/baseball (Kneidinger et al., 2001), but it was not
detectable in the basketball study of Kraus et al. (2010) nor in the
present study.
Looking at the remaining touch behaviors in the coding scheme, it
appears that the three behaviors low five, high five and high ten seem
to be behaviors generally used in a team sport setting: they appeared in
the current coding scheme, but also in the studies of Kneidinger et al.
(2001) and Kraus et al. (2010). However, two touching behaviors seem to
be idiosyncratic in the handball context: touching a teammates
shoulders, and giving a teammate a double touch emerged as behaviors
from this study, but are not mentioned in the other studies. Conversely,
the other two studies included far more touching behavior than our
study, indicating that studying other sports, and/or other time-frames
within a sport can lead to different coding schemes. The study by
Kneidinger et al. (2001) proposed 32 touching behaviors, whereupon 22
appeared with a frequency of greater than five. However,
softball/baseball matches are held in a completely different setting,
with many more possibilities for the athletes to display touch behaviors
than in the rush of a handball players' return to defense. Finally,
the study of Kraus et al. (2010) in basketball included 12 behaviors
that were shown when celebrating positive play.
The results of the study, as well as related studies described
earlier, reveal that touching behavior is one channel of nonverbal
behavior that is frequently seen within a team sport setting. Different
authors state that touch is related to the expression of emotions
(Richmond et al., 2012), and to specific positive emotions such as
happiness (Hertenstein et al., 2009; Hertenstein et al., 2006). However,
of importance from a team perspective is the assumption that touch can
function as positive reinforcement, foster intimacy, encourage
compliance, communicate liking, and be vital to trust, cooperation and
group functioning (Hertenstein et al., 2006; Kraus et al., 2010).
In general, the proposed coding scheme showed an adequate level of
reliability. Four of the eleven final behaviors (one fist down, two
fists down, touch shoulder, and double touch) resulted in an agreement
below the original threshold of 75%; however, these behaviors were found
to be especially sensitive to cross-coding with other, closely related
categories (e.g., one fist down and one fist up), which lead to the
decision to leave them in the coding scheme. The reliability of the
touch behaviors of the current study are comparable with the results of
Kraus et al. (2010) and Kneidinger et al. (2001). Interestingly, the
agreement in the intra-observer analyses was generally slightly higher
than in the inter-observer analyses. This supports existing literature
postulating that intra-observer agreement is systematically higher than
inter-observer agreement when measuring dynamic behaviors (Boyce et al.,
2000).
Even though interesting results emerged from this study, two
limitations of the proposed coding scheme need to be acknowledged.
First, the coding scheme only takes into account if a specific nonverbal
behavior is shown, but no measure of magnitude or intensity of the
behavior, as suggested by Miltenberger and Weil (2013), was captured.
Likewise, the duration of the behavior was not measured. From an
emotional contagion perspective, though, this may be of concern because
it is proposed that the same expression shown with greater intensity and
visibility leads to more contagion to those around because of the
greater attention they attract (Barsade, 2002). Similarly, it can be
hypothesized that the intensity and duration of the behavior can have an
impact on the formation of an impression in an opponent. However, as
handball is executed on a small field (compared with e.g., baseball or
soccer), it can be assumed that the expression of any emotion is likely
to reach the teammates or opponents. Further knowledge about the impact
of length and intensity of celebration behaviors is needed to judge the
impact of that limitation. Second, the present coding scheme does not
allow generalizations to be made outside the study population (i.e.,
female elite handball players). Further research should investigate how
applicable the present coding scheme is in other sports, and/or with
male athletes. The authors expect that for sports similar to handball
(e.g., in terms of size of the court, playing rules) such as basketball
or floorball, many of the proposed behaviors from the current coding
scheme could be adapted, but some other, more sport-specific nonverbal
behaviors may need to be added. Likewise, we would not expect a coding
scheme for male handball players to look completely different, but it
eventually differs in some few nonverbal behaviors.
Another point of concern that has to be considered is that
nonverbal communication, in contrast to verbal communication, does not
offer distinct meaning for a specific behavior (Riggio & Riggio,
2012). The discussion thus far about possible meanings and causes of
different nonverbal behaviors can be considered as a hypothesis, but the
exact meaning of, and reasons for the display of a specific nonverbal
behavior, cannot be found unless one asks the involved player in a
specific situation. However, regardless of that, the shown nonverbal
behavior actually forms the information the environment (i.e., teammates
or opponents) receives and reacts on. Thus, when research about the
impact of such nonverbal behaviors in a social context such as team
sports is done, the true meaning of the behavior for the individual is
not of so much importance.
Having developed this coding scheme, it is now available for other
research teams who are interested in verifying the utility and
reliability of this coding scheme, which would in crease confidence in
using the coding scheme in future research. Ultimately, the most
interesting step is to start investigating the extent to which elite
handball players show nonverbal behaviors in the post-shot period, and
under which circumstances they show more or less of these behaviors.
Thereafter, the decisive question to investigate is if nonverbal
behaviors in the post-shot period during a handball match have a
meaningful impact on team performance, an impact that could happen
through internal feedback loops (e.g., Price et al., 2012), emotional
contagion within the team (e.g., Hatfield et al., 1994), or impression
formation by opponents (e.g., Furley et al., 2012). If such a
relationship can be proven, this would offer many new avenues for sport
psychology consultants to design interventions aiming at optimizing
nonverbal behavior in team athletes.
References
Adler, P. A., & Adler, P. (1994). Observational Techniques. In
N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of Qualitative
Research. Thousand Oates, CA: SAGE.
Anderton, C. H., & Heckel, R. V. (1985). Touching behaviors of
winners and losers in swimming races. Perceptual and Motor Skills, 60,
289-290. doi: 10.2466/pms. 1985.60.1.289
App, B., McIntosh, D. N., Reed, C. L., & Hertenstein, M. J.
(2011). Nonverbal channel use in communication of emotion: How may
depend on why. Emotion, 11, 603-617. doi: 10.1037/a0023164
Baesler, E. J., & Burgoon, J. K. (1987). Measurement and
reliability of nonverbal behavior. Journal of Nonverbal Behavior, 11,
205-233. doi: 10.1007/BF00987254
Bandura, A. (1977). Social learning theory. Englewood Cliffs, N.J.:
Prentice Hall.
Barsade, S. G. (2002). The Ripple Effect: Emotional Contagion and
Its Influence on Group Behavior. Administrative Science Quarterly, 47,
644-675. doi: 10.2307/3094912
Bornstein, G., & Goldschmidt, C. (2008). Post-scoring behaviour
and team success in football. . In P. Andersson, P. Ayton & C.
Schmidt (Eds.), Myths and facts about football: The economics and
psychology of the world s greatest sport (pp. 113-123). Newcastle upon
Tyne: Cambridge Scholars Publishing.
Boyce, T. E., Carter, N., & Neboschick, H. (2000). An
evaluation of intraobserver reliability versus interobserver agreement.
European Journal of Behavior Analysis, 1, 107-114.
Brewer, C. J., & Jones, R. L. (2002). A five-stage process for
establishing contextually valid systematic observation instruments: The
case of the Rugby Union. The Sport Psychologist, 16, 138-159.
Brody, L. R., & Hall, J. A. (2008). Gender and emotion in
context. In M. Lewis, J. M. Haviland-Jones & L. Feldman Barrett
(Eds.), Handbook of emotions (3rd ed., pp. 395408). New York: Guilford
Press.
Bull, P., & Doody John, P. (2013). Gesture and body movement.
In J. A. Hall & M. L. Knapp (Eds.), Nonverbal Communication (pp.
205-227). Berlin: Mouton de Gruyter.
Burgoon, J. K., Buller, D. B., & Woodall, W. G. (1996).
Nonverbal communication: The unspoken dialog. New York: McGraw-Hill.
Cantrill, J. A., Sibbald, B., & Buetow, S. (1996). The Delphi
and nominal group techniques in health services research. The
International Journal of Pharmacy Practice, 4, 67-74. doi:
10.111/j.2042-7174.1996.tb00844.x
Cerin, E., Szabo, A., Hunt, N., & Williams, C. (2000). Temporal
patterning of competitive emotions: A critical review. Journal of Sports
Sciences, 18, 605-625.
Cohn, J. F., & Ekman, P. (2005). Measuring facial action. In J.
A. Harrigan, R. Rosenthal & K. R. Scherer (Eds.), The new handbook
of methods in nonverbal behavior research. Oxford: University Press.
Cone, J. D. (1977). The relevance of reliability and validity for
behavioral assessment. Behavior Therapy, 8,411-426. doi:
10.1016/S0005-7894(77)80077-4
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied
behavior analysis (2nd ed.). Upper Saddle River, N.J.: Pearson
Dael, N., Mortillaro, M., & Scherer, K. R. (2012a). Emotion
expression in body action and posture. Emotion, 12, 1085-1101. doi:
10.1037/a0025737
Dael, N., Mortillaro, M., & Scherer, K. R. (2012b). The Body
Action and Posture coding system (BAP): Development and reliability.
Journal of Nonverbal Behavior, 36, 97-121.
Darwin, C. (1872). The expression of the emotions in man and
animals. New York: Oxford University Press.
Ekman, P. (2003). Emotions revealed. New York: Times Books.
Ekman, P., & Friesen, W. V. (1969). The repertoire of nonverbal
behavior: Categories, origins, usage, and coding. Semiotica, 1, 49-98.
Ekman, R, & Friesen, W. V. (1972). Hand movements. The Journal
of Communication, 22, 353-374.
Ekman, P., & Rosenberg, E. L. (2005). What the face reveals:
basic and applied studies of spontaneous expression using the facial
action coding system (FACS). New York: Oxford University Press.
Fransen, K., Vanbeselaere, N., Exadaktylos, V., Vande Broek, G., De
Cuyper, B., Berckmans, D., ... Boen, F. (2012). "Yes, we
can!": Perceptions of collective efficacy sources in volleyball.
Journal of Sports Sciences, 30, 641-649.
Friesen, W. V., Ekman, R, & Wallbott, H. (1979). Measuring hand
movements. Journal of Nonverbal Behavior, 4, 97-112. doi:
10.1007/bf01006354
Furley, P., & Dicks, M. (2012). 'Hold your head
high'. The influence of emotional versus neutral nonverbal
expressions of dominance and submissiveness in baseball. International
Journal of Sport Psychology, 43, 294-311.
Furley, P., Dicks, M., & Memmert, D. (2012). Nonverbal behavior
in soccer: The influence of dominant and submissive body language on the
impression formation and expectancy of success of soccer players.
Journal of Sport & Excercise Psychology, 34, 61-82.
Furr, R. M., & Funder, D. C. (2007). Behavioral Observation. In
R. W. Robins, R. C. Realey & R. F. Krueger (Eds.), Handbook of
Research Methods in Personality Psychology (pp. 273-291). New York: The
Guilford Press.
Gee, C., & Sullivan, P. (2006). Using a direct observation
approach to study aggressive behaviour in hockey: Some preliminary
findings. Athletic Insight: The Online Journal of Sport Psychology, 8,
16-30.
Greenlees, I., Bradley, A., Holder, T., & Thelwell, R. (2005).
The impact of opponents' non-verbal behaviour on the first
impressions and outcome expectations of table-tennis players. Psychology
of Sport and Exercise, 6, 103-115. doi: 10.1016/j.
psychsport.2003.10.002
Greenlees, I., Buscombe, R., Thelwell, R., Holder, T., &
Rimmel, M. (2005). Impact of opponents' clothing and body language
on impression formation and outcome expectations. Journal of Sport &
Excercise Psychology, 27, 39-52.
Hall, J. A., & Gunnery, S. D. (2013). Gender differences in
nonverbal communication. In A. Hall Judith & L. Knapp Mark (Eds.),
Nonverbal Communication (pp. 639-670). Berlin: Mouton de Gruyter.
Harrigan, J. A. (2005). Proxemics, Kinesics, and Gaze. In J. A.
Harrigan, R. Rosenthal & K. R. Scherer (Eds.), The new handbook of
Methods in Nonverbal Behavior Research (pp. 137-198). Oxford: Oxford.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994).
Emotional Contagion. Paris: Cambridge University Press.
Heckel, R. V. (1993). Comparison of touching behavior of winners
and losers in racquetball and tennis. Perceptual and Motor Skills, 77,
1392-1394. doi: 10.2466/ pms.1993.77.3f.1392
Heckel, R. V., Allen, S. S., & Blackmon, D. C. (1986). Tactile
communication of winners in flag football. Perceptual and Motor Skills,
63, 553-554. doi: 10.2466/pms. 1986.63.2.553
Hertenstein, M. J., Holmes, R., McCullough, M., & Keltner, D.
(2009). The communication of emotion via touch. Emotion, 9, 566-573.
doi: 10.1037/a0016108
Hertenstein, M. J., Verkamp, J. M., Kerestes, A. M., & Holmes,
R. M. (2006). The Communicative Functions of Touch in Humans, Nonhuman
Primates, and Rats: A Review and Synthesis of the Empirical Research.
Genetic, Social, and General Psychology Monographs, 132, 5-94. doi:
10.3200/mono.132.1.5-94
Jones, J., & Hunter, D. (1995). Consensus methods for medical
and health services research. BMJ, 311, 376-380.
Jones, M., & Harwood, C. (2008). Psychological momentum within
competitive soccer: Players' perspectives. Journal of Applied Sport
Psychology, 20, 57-72. doi: 10.1080/10413200701784841
Kappas, A., Krumhuber, E., & Kiister, D. (2013). Facial
behavior. In J. A. Hall & M. L. Knapp (Eds.), Nonverbal
Communication (pp. 131-165). Berlin: Mouton de Gruyter.
Kazdin, A. E. (2010). Single case research designs: Methods for
clinical and applied settings (2nd ed.). New York: Oxford University
Press.
Kneidinger, L. M., Maple, T. L., & Tross, S. A. (2001).
Touching behavior in sport: Functional components, analysis of sex
differences, and etholocigal considerations. Journal of Sport Behavior,
25, 45-62. doi: 10.1023/A:1006785107778
Kraus, M. W., & Chen, T.-W. D. (2013). A winning smile? smile
intensity, physical dominance, and fighter performance. Emotion, doi:
10.1037/a0030745
Kraus, M. W., Huang, C., & Keltner, D. (2010). Tactile
communication, cooperation, and performance: An ethnological study of
the NBA. Emotion, 10, 745-749. doi: 10.1037/ a0019382
Lazarus, R. S. (1999). Stress and emotion: a new synthesis. New
York: Springer Publishing Co.
Matsumoto, D., & Hwang, H. S. (2012). Evidence for a nonverbal
expression of triumph. Evolution and Human Behavior, 33, 520-529. doi:
10.1016/j.evolhumbehav.2012.01.005
Matsumoto, D., & Willingham, B. (2006). The thrill of victory
and the agony of defeat: Spontaneous expressions of medal winners of the
2004 Athens Olympic games. Journal of Personality and Social Psychology,
91, 568-581. doi: 10.1037/0022-3514.91.3.568
Mellalieu, S. D., Neil, R., Hanton, S., & Fletcher, D. (2009).
Competition stress in sport performers: Stressors experienced in the
competition environment. Journal of Sports Sciences, 27(1), 729-744.
doi: 10.1080/02640410902889834
Miltenberger, R. G., & Weil, T. M. (2013). Observation and
measurement in behavior analysis. In G. J. Madden, W. V. Dube, T. D.
Hackenberg, G. P. Hanley & K. A. Lattal (Eds.), APA handbook of
behavior analysis, Vol. 1: Methods and principles (pp. 121-150).
Washington, DC, US: American Psychological Association.
Mitchell, S. K. (1979). Interobserver agreement, reliability, and
generalizability of data collected in observational studies.
Psychological Bulletin, 86, 376-390. doi: 10.1037/00332909.86.2.376
Moesch, K., & Apitzsch, E. (2012). How do coaches experience
Psychological Momentum? A qualitative study of female elite handball
teams. The Sport Psychologist, 26, 435-453.
Moll, T., Jordet, G., & Pepping, J.-G. (2010). Emotional
contagion in soccer penalty shootouts: Celebration of individual success
is associated with ultimate team success. Journal of Sport Sciences, 28,
983-992. doi: 10.1080/02640414.2010.484068
Price, T. F., Peterson, C. K., & Harmon-Jones, E. (2012). The
emotive neuroscience of embodiment. Motivation and Emotion, 36, 27-37.
doi: 10.1007/sl 1031-011-9258-1
Richmond, V. P., McCroskey, J. C., & Hickson III, M. L. (2012).
Nonverbal behavior in interpersonal relations (7th ed.). Allyn &
Bacon: Pearson.
Riggio, R. E., & Riggio, H. R. (2012). Face and body in motion:
Nonverbal communication. In T. F. Cash (Ed.), Encyclopedia of body image
and human appearance (Vol. 1). (pp. 425-430). San Diego, CA US: Elsevier
Academic Press.
Ronglan, L. T. (2007). Building and communicating collective
efficacy: A season-long in-depth study of an elite sport team. The Sport
Psychologist, 21, 78-93.
Rosenthal, R. (2005). Conducting judgment studies: Some
methodological issues. In J. A. Harrigan, R. Rosenthal & K. R.
Scherer (Eds.), The new handbook of methods in nonverbal behavior
research (pp. 199-236). New York: Oxford University Press.
Shariff, A. F., & Tracy, J. L. (2011). What Are Emotion
Expressions For? Current Directions in Psychological Science, 20,
395-399. doi: 10.1177/0963721411424739
Sherif, M. (1966). Group conflict and co-operation: Their social
psychology. London: Routledge & Regan Paul.
Stangor, C. (2007). Research Methods for behavioral Sciences (3
ed). Boston: Houghton Mifflin.
Totterdell, P. (2000). Catching moods and hitting runs: Mood
linkage and subjective performance in professional sport teams. Journal
of Applied Psychology, 85, 848-859. doi: 10.1037/0021-9010.85.6.848
Tracy, J. L., & Matsumoto, D. (2008). The spontaneous
expression of pride and shame: Evidence for biologically innate
nonverbal displays. Proceedings of the National Academy of Sciences,
105, 11655-11660. doi: 10.1073/pnas.0802686105
Tracy, J. L., & Robins, R. W. (2004a). Putting the self into
self-conscious emotions: A theoretical model. Psychological Inquity, 15,
103-125. doi: 10.1207/s15327965plil502_01
Tracy, J. L., & Robins, R. W. (2004b). Show your pride:
evidence for a discrete emotion expression. Psychological Science, 15,
194-197. doi: 10.1111/j.09567976.2004.01503008.x
Tracy, J. L., & Robins, R. W. (2007a). The prototypical pride
expression: Development of a nonverbal behavior coding system. Emotion,
7, 789-801. doi: 10.1037/1528-3542.7.4.789
Wallbott, H. G. (1998). Bodily expression of emotion. European
Journal of Social Psychology, 28, 879-896. doi:
10.1002/(SICI)1099-0992(1998110)28:6<879::AIDEJSP901>3.0.CO;2-W
Van de Ven, A. H., & Delbecq, A. L. (1972). The nominal group
as a research instrument for exploratory health studies. American
Journal of Public Health, 62, 337-342.
Warr, P. B., & Knapper, C. (1968). The perception of people and
events. London: John Wiley & Sons.
Watkins, M. W., & Pacheco, M. (2000). Interobserver agreement
in behavior research: Importance and calculation. Journal of Behavioral
Education, 10, 205-212.
Yoder, P., & Symons, F. (2010). Observational measurement of
behavior. New York: Springer.
Karin Moesch
Lund University
Goran Kentta and C. Mikael Mattsson
The Swedish School of Sport and Health Sciences (GIH), Stockholm
Address correspondence to: Karin Moesch, Department of Psychology,
Lund University, Box 213, 22 100 Lund, Sweden. Email:
karin.moesch@psy.lu.se
Table 1a
Preliminary coding scheme for positive PSB
Category Behavior Description
PSB-P One fist down One hand in a fist, arm bent,
(Positive fist at a highest at the level of
PSB) the shoulders, before or beside
the body
Two fists down Two hands in fists, arms bent,
fists at a highest at the level
of the shoulders, before or
beside the body
One fist up One hand in a fist, arm bent or
stretched, at the height of the
neck or higher, before or beside
the body
Two fists up Two hands in fists, arms bent or
stretched, at the height of the
neck or higher, before or beside
the body
One arm up One arm up, without having a fist
Two arms up Two arms up, without having a
fist
Airplane Two arms stretched at the side of
the body, horizon- tally, arms
completely stretched or max. a
little bent
Smile * Smile
Laugh * Laugh
Erected body posture * Erected body posture
High tempo Player moves very quickly, which
is, however, not dependent on the
game
Thumbs up Fist shown at some place and the
thumb is showing up
Fighting for the ball Player engages strongly in
getting the ball, but not in an
aggressive way
Clapping hands Clapping in the hands
Note. * = Behaviors that were deleted in the revised version of the
coding scheme.
Table 1b
Preliminary coding scheme for negative PSB
Category Behavior Description
PSB-N Slumped body posture * Slumped body posture
(Negative Kick into the bench Run back
PSB) Low tempo slower than the game process
would expect, or stand or lie
still before moving back to
defense
Kicking in the bench * Kick into the bench
Kicking in the floor * Kick into the floor
Displaying expressions Code for all expressions of
of frustration/irritation frustration/irritation (e.g.,
a very aggressive defense
leading to a two-minute
penalty, or throwing the arms
as a sign of frustration)
Note. * = Behaviors that were deleted in the revised version of the
coding scheme.
Table 1c
Preliminary coding scheme for touch
Category Behavior Description
PSB-T Low five Touch hand to hand (or fist to fist)
(PSB that happens up to the height of the
chest
Touch) High five Touch hand to hand (or fist to fist)
that happens above the height of the
chest
Low ten Touch two hands to two hands (or
fists to fists) that happens up to
the height of the chest
High ten Touch two hands to two hands (or
fists to fists) that happens above
the height of the chest
Touch bottom Touch on the bottom or the lower
Hug * part of the back Hug
Touch shoulders Touch on the shoulders or the higher
part of the back
Double touch Touch containing two different
touches at two different places
simultaneously (e.g., low five and a
touch on the shoulders)
Note. * = Behavior that was deleted in the revised version of the
coding scheme.
Table 2
Results of the inter-observer and intra-observer agreement analysis
for specific behaviors of the final version of the coding scheme
Revised coding scheme
Category Behavior % inter-OA % intra-OA
PSB-P (Positive One fist down 63% (112) 70% (30)
PSB) Two fists down 67% (66) 71% (21)
One fist up 79% (159) 82% (45)
Two fists up 82% (155) 83% (42)
One arm up 0%(1) (0)
Two arms up 50% (6) (0)
Airplane 47% (17) 100% (1)
High tempo 0% (5) 0%(1)
Thumbs up 71% (17) 100% (4)
Fighting for the 0% (1) 100% (1)
ball
Clapping hands 87% (15) 75% (8)
PSB-N (Negative Low tempo 56% (9) 20% (5)
PSB) Displaying expressions 40% (15) 100% (7)
of frustration/
irritation
PSB-T (PSB Low five 84% (179) 90% (40)
touch) High five 94% (190) 97% (38)
Low ten 50% (2) (0)
High ten 94% (65) 100% (19)
Touch bottom 50% (2) 0% (2)
Touch shoulders 57% (14) 75% (4)
Double touch 57% (7) 80% (5)
Final coding scheme
Category Behavior % inter-OA % intra-OA
PSB-P (Positive One fist down 63% (112) 70% (30)
PSB) Two fists down 67% (66) 71% (21)
One fist up 79% (159) 82% (45)
Two fists up 88% (164) * 84% (43) *
One arm up na na
Two arms up na na
Airplane * *
High tempo na na
Thumbs up 71% (17) 100% (4)
Fighting for the na na
ball
Clapping hands 87% (15) 75% (8)
PSB-N (Negative Low tempo na na
PSB) Displaying expressions na na
of frustration/
irritation
PSB-T (PSB Low five 84% (179) 90% (40)
touch) High five 94% (190) 97% (38)
Low ten na na
High ten 94% (65) 100% (19)
Touch bottom na na
Touch shoulders 57% (14) 75% (4)
Double touch 57% (7) 80% (5)
Note. OA = observer agreement; in parentheses is the total number
of behaviors that appeared; * in the final analyses, the
behaviors airplane and two fists up were merged; na: those
behaviors were deleted in the final version due to low
reliability.
Table 3
Results of the inter-observer and intra-observer agreement
analyses of the final version of the coding scheme (Handball
Post-Shot Behavior Coding Scheme, H-PSB-CS)
% inter-observer % intra-observer
Category agreement agreement
PSB-G (PSB gesture) 84% (490) 86% (138)
PSB-T (PSB touch) 91% (439) 95% (100)
Overall 87% (927) 90% (238)
Mean count per interval 85% (345) 88% (84)
IOA
Note. In parentheses is the total number of behaviors that appeared
Table 4
Final version of the coding scheme (Handball Post-Shot Behavior
Coding Scheme, H-PSB-CS)
Category Behavior
Post-shot behavior gesture (PSB-G) One fist down
Two fists down
One fist up
Two fists up
Thumbs up
Clapping hands
Post-shot behavior touch (PSB-T) Low five
High five
High ten
Touch shoulders
Double touch
