Sport & Exercise Psychology Review, Vol. 5 No. 2

3

© The British Psychological Society 2009

RAWLEY AND PASKEVICH (1997)
defined team building as ‘a method of
helping the group to: (a) increase effec-

tiveness; (b) satisfy the needs of its members;
or (c) improve work conditions’ (p.13). In
short, team building interventions are used to
promote and enhance the functioning of a
group (Beer, 1980; Carron & Hausenblas,
1998) – to help a group become more cohe-
sive while working toward its common goals
and objectives. Essentially, four approaches
have been taken; either singly or in combina-
tion, team building interventions have focused
on role relationships, interpersonal relation-
ships, goal setting or leadership (Beer, 1980).

Team building has been used with a large

variety of groups in a number of areas of
study. One of those areas has been organisa-
tional psychology with work teams. In one
study that is illustrative, Buller and Bell
(1986) employed team building and goal set-
ting strategies with a group of 53 hard-rock
miners in an attempt to enhance productiv-
ity and strategy development. Although
slight improvements were found in some
measures of performance, the overall results
were inconclusive. As another example,
Eden (1986) had seven military companies
undergo a three-day team development
workshop. In comparison to nine control
combat companies, companies exposed to
the intervention significantly improved in
teamwork, conflict handling, and under-
standing about plans.

In an attempt to introduce coherence

into a literature characterised by inconsis-
tent findings, summaries of the available
research have been undertaken. For exam-
ple, in 1989, Neuman et al. carried out a
meta-analysis of 126 studies in business and
industry. These studies employed a wide vari-
ety of interventions (not just team building)
designed to improve attitudes of individuals
towards their jobs, their organisations, and
even themselves. The results showed that
team building was the most effective inter-
vention for improving individual satisfaction
and changing attitudes.

A second area of study where team build-

ing interventions have been used is exercise
psychology with physical activity groups/
classes. Generally, the two primary outcomes
targeted in this literature have been the level
of cohesiveness within the exercise classes and
improved adherence of individual members.
In one study, Carron and Spink (1993)
assessed the perceptions of cohesion and rat-
ings of satisfaction of 94 female university stu-
dents enrolled in a standard exercise class that
met three times per week for 13 weeks. It was
found that in comparison to participants in
control (i.e. ‘standard’) exercise classes, par-
ticipants in classes exposed to a team building
intervention had stronger perceptions of
cohesion (specifically, the Individual Attrac-
tion to the Group-Task [ATG-T] dimension),
as well as increased individual satisfaction and
satisfaction with the fitness class experience.

Team building interventions in sport:
A meta-analysis

Luc J. Martin, Albert V. Carron & Shauna M. Burke

A meta-analysis of team building interventions in sport was completed. Seventeen studies containing
180 effect sizes were retrieved. The overall effect (Hedges g) was .427. Analyses of possible moderator
variables showed the largest effect sizes were in interventions where: (a) non-experimental designs were used
(g=.474); (b) the data were unpublished (g=.539); (c) goal setting only was used (g=.714); (d) the
coach/manager directed the delivery (g=.446); and (e) the teams were at the university level (g=.482).
Finally, team building had the greatest influence on cognitions (g=.799).

B

As another example, Estabrooks and Car-

ron (1999) examined the effectiveness of a
team building intervention on class cohesion,
individual adherence, and return rates 10
weeks following cessation of the exercise pro-
gramme. The participants were 33 older
adults (mean age=75.1) who were divided into
three groups: team building, placebo, and
control. The results showed that participants
in the team building condition attended more
classes (90.8 per cent) and had a higher
return rate (91.7 per cent) than participants in
both the control (65 per cent and 40 per cent,
respectively) and placebo (70 per cent and 73
per cent, respectively) conditions.

The influence of team building in exer-

cise classes also has been summarised using
meta-analysis. In 2006, Burke et al. empiri-
cally compared the results from 44 studies
that employed four different physical activity
intervention contexts: home-based pro-
grammes without contact from researchers
or health care professionals, home-based
programmes with contact from researchers
or health care professionals, standard exer-
cise class programmes, and team building
programmes (i.e. studies where the exercise
class was introduced to a team building inter-
vention designed to approximate it to a true
group). The results showed that in terms of
adherence, exercising in a true group was
superior to exercising in a standard exercise
class, exercising at home with contact or
exercising at home without contact.

A third area of study where team building

interventions have been used is sport. Gener-
ally, the two primary outcomes targeted have
been intragroup relationships (e.g. cohe-
sion) and performance (e.g. team success).
For example, in a study carried out by
McClure and Foster (1991), 15 university
gymnasts were subdivided into an interven-
tion (N=8) and a control (N=7) condition.
The intervention group met weekly for 15
weeks to explore personal issues relevant to
personal and team development. Also,
opportunities were provided to explore feel-
ings, resolve inter-member conflicts, and
address areas of personal concern outside of

the team environment. The authors
reported that athletes in the intervention
condition moved toward increased levels of
cohesion and had a higher attraction to the
group than team members in the control
condition.

As another example, Prapavessis et al.

(1996) examined the influence of a team
building intervention programme on cohe-
sion and team success in soccer teams. Seven
First Division teams from the Western
Ontario soccer league were subdivided into
three groups: team building (N=three teams,
50 athletes), placebo-control (N=two teams,
49 athletes), and control (N=two teams, 28
athletes). No significant differences were
found in performance or cohesion among
the three conditions.

The two studies just cited serve to high-

light some of the difficulties and resulting
limitations in research in sport psychology
that has attempted to empirically examine
the effectiveness of a team building interven-
tion. One limitation is associated with power.
Team building is a group-based intervention
and obtaining a sufficient number of teams
to carry out a well-controlled study is diffi-
cult. Securing the co-operation of a whole
league, for example, is a remarkable recruit-
ing achievement; from a scientific perspec-
tive, however, having sample sizes of only two
to four teams in the experimental, control,
and placebo conditions is a limitation.

A second limitation, one that likely flows

from the first, is that a substantial number of
team building studies in sport psychology
employ a single team, pre-post, non-experi-
mental design. The lack of a control condition
(and the relatively small number of partici-
pants within the intervention condition)
makes any conclusion(s) drawn about the
effectiveness of team building problematic.

A third limitation, again one that likely is

associated with a lack of power (and in some
cases, the use of single team pre-post
designs), is that a considerable number of
theses and dissertations that have focused on
team building in sport have not been pub-
lished. A body of unpublished research typi-

4

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

cally has less impact on our understanding of
any area of knowledge. It is difficult to deter-
mine the reason underlying the relatively
large body of unpublished team building
studies in sport psychology. It could be, for
example, a lack of enthusiasm on the part of
the researcher to prepare a manuscript after
(finally) completing a thesis/ dissertation. It
could also be that journal reviewers and edi-
tors were unwilling to publish research lack-
ing power and/or control groups. Whatever
the reason(s), this substantial body of
research has the potential to inform the sci-
ence and practice of sport psychology about
team building.

Meta-analysis offers two important advan-

tages for the summary of a body of literature;
those advantages have particular relevance
to the question of the effects of team build-
ing in sport. First, any meta-analysis results in
an increase in power over single (some-
times) inconclusive studies. Second, meta-
analysis provides for the opportunity to test
empirically the influence of suspected mod-
erator variables such as, for example, possi-
ble differences in the results from published
and unpublished studies.

Thus, the main purpose of the present

study was to conduct a meta-analysis to exam-
ine the effect of team building interventions
in sport. As a secondary purpose, the influ-
ence of a number of potential moderator
variables also was examined. The specific
variables examined are outlined in the sec-
tion which follows.

Characteristics of the study

Experimental design. As was pointed out
above, a substantial number of team build-
ing studies in sport psychology have
employed a single team, pre-post, non-exper-
imental design; others have used a quasi-
experimental design. Consequently, one
question of interest in the present study was
to examine whether results from studies
using non-experimental designs differ from
those using a quasi-experimental design
(i.e. possess both experimental and control
conditions).

Sources of data. Over 40 years ago, Rosenthal
(1966) spoke to the propensity for agencies
(e.g. doctoral committees, journal editorial
boards) to favour studies with significant
findings. Thus, meta-analyses (including the
present one) routinely compare effect sizes
from refereed publications versus non-pub-
lished sources.

Characteristics of the team building
intervention

Intervention types. The term ‘team building’ is
an umbrella-like term that has been used to
describe a large number of interventions in
industrial psychology (Beer, 1980; Blake &
Mouton, 1964). Not surprisingly, perhaps,
the same situation exists in sport psychology.
For example, team building interventions
with sport teams have consisted of: (a) a sin-
gle task-oriented protocol such as goal-set-
ting (e.g. Burton, 1989; Stevens & Bloom,
2003); (b) an omnibus task-oriented proto-
col that incorporates a wide variety of psy-
chological constructs (e.g. Prapavessis et al.,
1996); (c) a socially-oriented protocol that is
directed toward interpersonal relations (e.g.
McClure & Foster, 1991); and (d) an adven-
ture/outdoor experience (e.g. Rainey &
Schweickert, 1988). Thus, a question of
interest in the present study was to deter-
mine whether there are differences in effec-
tiveness among the various types of
interventions employed for team building in
sport.

Mode of intervention delivery. Carron and
Hausenblas (1998) pointed out that two
types of protocols have been used to deliver
team building interventions in sport. In one,
the direct approach, a consultant (e.g. sport
psychologist) introduces and implements
the intervention with the athletes/team. In
the second, the indirect approach, the con-
sultant works with the coach or team man-
ager who then implements the programme
developed. In the latter case, the team build-
ing intervention is filtered through the
coach/manager. A question of interest
examined in the present study was whether

Sport & Exercise Psychology Review, Vol. 5 No. 2

5

Team building interventions in sport: A meta-analysis

the direct approach for delivering a team
building intervention is more effective than
the indirect approach.

Intervention length. In their literature review
of team building interventions in organisa-
tional psychology, DeMeuse and Liebowitz
(1981) proposed that a serious limitation was
study duration. Brawley and Paskevich
(1997) raised this same concern for team
building research in sport, noting that ‘it
seems that the proper evaluation of any
[team building] intervention would require
a minimum of a season for any meaningful,
enduring changes to be validly assessed’
(p.20). Therefore, a question of interest per-
taining to the characteristics of the interven-
tion was whether the length of the
intervention with a sport team has an impact
on its effectiveness.

Characteristics of the participants

Gender. Consistently, gender has been shown
to be a moderator variable across psycholog-
ical domains including, for example, com-
petitive orientation (Duda, 1987; Gill, 1986),
and emergence into task and social leader-
ship roles (Eagly & Karau, 1991). Given that
competitive sport is not coeducational and
that team building interventions have been
undertaken with both male and female
teams, another question of interest was con-
cerned with whether gender is a moderator
in the team building-team effectiveness rela-
tionship.

Type of sport. In their meta-analysis, Carron et
al. (2002) found support for a conclusion
that sport type (categorised as interactive
versus coactive) is a moderator in the cohe-
sion-performance relationship (i.e. the cohe-
sion-performance relationship is stronger in
coactive sports). Carron et al. (2002) pro-
posed that team building interventions may
have a greater impact in coactive sports as
there are fewer opportunities for group
identity to develop. Therefore, a question of
interest was whether team building has a dif-
ferent impact in different types of sport.

Skill level. In Chelladurai’s (1978) multidi-
mensional model of leadership, the category
member characteristics is proposed to be an
important moderator of the relationship
between coaching behaviour and athlete
performance and satisfaction; skill level is a
member characteristic. Consequently, a
question of interest was whether team build-
ing has a different impact on sport teams of
different skill level.

Dependent variables

Cohesiveness. Cohesion has been defined by
Carron and Hausenblas (1998) as ‘a dynamic
process that is reflected in the tendency for a
group to stick together and remain united in
the pursuit of its instrumental objectives
and/or for the satisfaction of member affec-
tive needs’ (p.213). Given the well docu-
mented relationships between both task and
social cohesion and performance (Carron et
al., 2002), it’s hardly surprising that a large
number of team building studies in sport
have used cohesion as a measure of the effec-
tiveness of an intervention. Thus, the first
two dependent variables examined in the
present meta-analysis were social and task
cohesion.

Performance. Outcomes of fundamental con-
cern in any task-oriented group are those
pertaining to individual and collective per-
formance (Carron et al., 2002). Thus,
another dependent variable examined in the
meta-analysis was performance.

Enhanced cognitions. Although the team
building literature in sport is not extensive, a
relatively wide variety of cognitions have
been assessed (e.g. perceptions of satisfac-
tion and quality of life). As a result of the rel-
ative lack of effect sizes associated with any
single cognition, two options were available –
disregard some data or combine the effect
sizes from various similar cognitions into one
category (the specific variables in the
enhanced cognitions category are outlined
below). Therefore, a fourth dependent vari-
able examined was enhanced cognitions.

6

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Roles. Woodcock and Francis (1994) have
suggested that when team building is suc-
cessful team members understand and
accept their roles and responsibilities. The
present study provided an opportunity to
empirically examine this suggestion; a fifth
dependent variable was team roles.

Anxiety. Baumeister and Leary (1995) pro-
posed that membership in a group positively
benefits cognitions such as anxiety. Also, Pra-
pavessis et al. (1996) reported that athletes
possessing higher perceptions of task cohe-
sion also reported less cognitive anxiety.
Some research in sport psychology has exam-
ined the influence of a team building inter-
vention on athlete anxiety. Consequently,
anxiety was the sixth dependent variable
examined.

Method

Selection of the studies for inclusion
Studies were obtained using computer
searches, manual searches, and journal
searches, as well as through personal corre-
spondence. The computer searches con-
sisted of an investigation of various
computer databases: PsychINFO (1806-cur-
rent), PsycARTICLES, SPORT Discus (1830-
current), and DISSERTATION ABSTRACTS
ONLINE. The keywords presented for the
computer searches included team building,
sport, interventions, goal setting, cohesion and
various combinations of these. The manual
searches involved utilising reference lists
from published articles and narrative reviews
(e.g. Bloom & Stevens, 2002; Brawley &
Paskevich, 1997; Dunn & Holt, 2004; Holt &
Dunn, 2006). The journal searches focused
on publications most likely to contain rele-
vant research with regard to team building
interventions in sport. The journals exam-
ined included Journal of Sport and Exercise Psy-
chology, Research Quarterly for Exercise and Sport,
Small Group Research, The Sport Psychologist,
and International Journal of Applied Sport Sci-
ences. The personal correspondence aspect
involved e-mail contact with researchers
active in the area of applied sport psychology

to determine if they were in possession of
unpublished data related to team building.

The various searches produced 32 poten-

tial articles and data sets containing team
building interventions. Given the nature of
the research question, all studies ultimately
included in the meta-analysis: (a) contained
a team building intervention with a sport
team; and (b) possessed sufficient statistical
information for the computation of an effect
size. The result was a total of 17 studies con-
taining 180 effect sizes.

Many of the original sample of 180 effect

sizes represented multiple endpoints (i.e.
multiple measures of the same dependent
and independent variables in a single study).
For example, the Kilty (2000) team building
study produced 12 effect sizes because the
Group Environment Questionnaire was
administered to assess perceptions of cohe-
sion among rowers at three different times
throughout the course of the season. From a
meta-analytic perspective, multiple end-
points violate the assumption of independ-
ent data points (Bangert-Drowns, 1986;
Gleser & Olkin, 1994). In short, the 12 effect
sizes in the Kilty study pertaining to a single
question in the present meta-analysis (i.e.
the influence of a team building interven-
tion on perceptions of group cohesion) had
the potential to exert an inordinate influ-
ence on the results. For this reason, in the
Kilty study and other similar studies, average
effect sizes were computed: (a) where multi-
ple endpoints were present; but (b) in a
manner that permitted examination of the
major questions of interest. The averaging of
effect sizes was undertaken after each effect
size was adjusted statistically as per the pre-
scriptions outlined by Hedges and Olkin
(1985; see below for further elaboration on
this point). The result was a final sample of
80 effect sizes.

Dependent variables

The 17 studies contained a variety of
dependent variables, many of which were
conceptually similar. Therefore, the depend-
ent variables were combined into the six gen-

Sport & Exercise Psychology Review, Vol. 5 No. 2

7

Team building interventions in sport: A meta-analysis

eral categories outlined in the Introduction:
social cohesion (attractions to the group-
social, group integration-social, attractions
to the group), task cohesion (attractions to
the group-task, group integration-task, unity
of purpose, quality of teamwork, athletic
communication), performance (self-
reported performance, coaches perceptions
of performance, concentration, effort, goal
setting accuracy, performance), enhanced
cognitions (satisfaction, self-confidence,
quality of life, expectancy accuracy, per-
ceived success, international success attribu-
tions, perceivability, self efficacy, hope), roles
(role clarity, role acceptance, role perform-
ance, valued roles), and anxiety (cognitive
anxiety, somatic anxiety).

Two researchers carried out the coding

for each study. The data were not transferred
from the original source to the coding sheets
until both researchers were in agreement
about categorisation aspects. Reliability was
further strengthened by having two
researchers present when the data were
entered into the computer file.

Independent variables

The main research question, as was pointed
out above, pertained to the relative effective-
ness of team building interventions in sport.
In the 17 studies included in the meta-analy-
sis, the nature of the team building interven-
tions varied. The interventions included goal
setting (collective or individual) which pro-
duced 22 effect sizes, adventure (outdoor)
programmes which produced 18 effect sizes,
interpersonal relations which produced 11
effect sizes, and omnibus (programmes where
a combination of team building strategies
were introduced) which produced 29 effect
sizes. The effectiveness of each of these was
examined.

Characteristics of the study

Experimental design. As indicated above, the
influence of a number of potential modera-
tor variables was examined. One moderator
variable examined was study design. In this
regard, a distinction was made between

studies employing non-experimental (i.e. no
control condition) versus quasi-experimental
(i.e. experimental and control conditions)
designs. Type of design produced 31 (non-
experimental) and 49 (quasi-experimental)
effect sizes respectively.

Source of data. Fifty-nine effect sizes were
obtained from refereed publications; 21 from
unpublished sources (e.g. theses).

Characteristics of the team building
intervention

Intervention type. As mentioned above, goal
setting interventions produced 22 effect sizes,
adventure programmes produced 18 effect
sizes, omnibus interventions produced 29 effect
sizes, and interpersonal relation interventions
produced 11 effect sizes.

Mode of delivery. Studies using a direct method,
where a consultant delivered the interven-
tion to the team, produced 39 effect sizes.
The indirect method, where the intervention
was delivered to the coach/manager, who in
turn delivered it to the team, produced 41
effect sizes.

Intervention length. The length of the inter-
vention varied markedly across the studies in
the sample and, therefore, was another pos-
sible moderator variable. Studies where the
length of the intervention was less than two
weeks produced 19 effect sizes, two weeks to 20
weeks produced 26 effect sizes, and finally, an
intervention length of 20 weeks and greater
produced 35 effect sizes. The division of
intervention length into these three cate-
gories was arbitrary – based on an objective
to obtain a relatively equal sample size across
categories.

Characteristics of the participants

Gender. Across the studies examined, three
different types of samples were used: males,
which produced 28 effect sizes, females,
which produced 47 effect sizes, and mixed
(male and female) samples, which produced
five effect sizes.

8

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Type of sport. The sports used in the interven-
tions were placed into three categories: Inter-
active, individual, or a combination. Interactive
(including, soccer, baseball, softball, basket-
ball, volleyball, hockey, rowing) sports pro-
duced 34 effect sizes, while individual
(including, gymnastics, track, equestrian,
swimming) and combination (a mix between
the previously mentioned sports) sports pro-
duced 41 and five effect sizes respectively.
The combination category resulted from
studies where a team building intervention
was administered to a variety of teams in dif-
ferent sport types.

Skill level. High school teams produced a total
of four effect sizes, while intercollegiate sports
produced 66 effect sizes and post-intercolle-
giate produced 10 effect sizes.

Analyses

Calculation of effect sizes. The calculation of
effect sizes was carried out using the pro-
gramme Comprehensive Meta-Analysis
(CMA, www.meta-analysis.com/index.html).
The programme offers a number of options
with respect to the type of effect size calcu-
lated. Hedges g (Hedges, 1981, 1982;
Hedges & Olkin, 1985) was selected – an
effect size that is adjusted to take into
account differences in sample size and sam-
ple variance.

The interpretation of effect size magni-

tude adhered to the prescriptions advanced
by Cohen (1988). That is, a Hedges g of .80
was interpreted to represent a large effect,
one of .50 a medium effect, and one of .20 a
small effect.

Meta-analytic model. It is possible to use either
a fixed effects or random effects model
(Field, 2001; Hunter & Schmidt, 2000) for
meta-analyses. In the present study, a ran-
dom effects model was used. One reason for
this is that research has shown that fixed
effects models have: (a) higher Type 1 error
rates; and (b) narrower confidence intervals
(Hunter & Schmidt, 2000). Field (2001) also
noted that if researchers wish to generalise

their findings with regards to effect size
parameters further than the set of observed
studies, a random effects approach is neces-
sary.

Tests for statistical significance. The signifi-
cance of an effect size can be estimated using
a variety of protocols. In the present study,
the 95 per cent confidence interval (CI

95

) was

used to determine if an effect size is signifi-
cantly different from zero. Thus, an effect
size is considered to be statistically significant
if CI

95

does not include or pass through the

value of zero (Hedges & Olkin, 1985).

To test for significant differences

between effect sizes among moderator vari-
ables, the effect sizes were rank ordered
using the SPSS programme. T-tests were
then conducted on all moderator variable
pairs. For those containing three or more
variables, a one-way Analysis of Variance
(ANOVA) was computed using the Bonfer-
roni procedure (Thomas et al., 1999) when
statistical significance was found (i.e. the L
value was greater than the tabled chi-square
value). The L statistic was found by dividing
the sum of the squares between, by the sum of the
squares within.

The Fail Safe N statistics were also calcu-

lated for all of the variables presented in
Table 2. The purpose is to test the extent that
journals have a tendency to only publish sta-
tistically significant findings (Rosenthal,
1979).

Results

Descriptive statistics
Table 1 contains the descriptive statistics per-
taining to the 17 studies included in the
meta-analysis. An examination of Table 1
shows that the studies in the meta-analysis
tended to be composed of females (58.8 per
cent), between 15 and 25 years of age (64.7
per cent), and on intercollegiate (82.4 per
cent), and interactive teams (58.8 per cent).
Further, the interventions analysed were
mostly from published sources (64.7 per
cent) using a direct style of delivery (64.7 per
cent).

Sport & Exercise Psychology Review, Vol. 5 No. 2

9

Team building interventions in sport: A meta-analysis

10

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Table 1: Descriptive statistics.

Study/Sample characteristics

Number (%) of studies

Design

Quasi-experimental

10 (58.8%)

Non-experimental 7

(41.2%)

Source

Published

12 (70.6%)

Unpublished

5 (29.4%)

Type of intervention

Goal setting

3 (17.6%)

Interpersonal relations

1 (5.9%)

Adventure programme

4 (23.5%)

Omnibus

9 (52.9%)

Delivery

Direct

11 (64.7%)

Indirect

6 (35.3%)

Length of intervention

Less than 2 weeks

5 (29.4%)

2 to 20 weeks

5 (29.4%)

20 weeks and above

7 (41.2%)

Gender

Male

5 (29.4%)

Female

10 (58.8%)

Mixed

2 (11.8%)

Sport type

Interactive

10 (58.8%)

Individual

6 (35.3%)

Combination

1 (5.9%)

Skill level

High School

1 (5.9%)

Intercollegiate

14 (82.4%)

Post-intercollegiate

2 (11.8%)

Dependent variables*

Social cohesion

11 (64.7%)

Task cohesion

11 (64.7%)

Performance

5 (29.4%)

Enhanced cognitions

4 (23.4%)

Roles

1 (5.9%)

Anxiety

3 (17.6%)

Mean age of participants

Less than 15 years

1 (5.9%)

15–25 years

11 (64.7%)

25–35 years

1 (5.9%)

35 years and greater

0 (0.0%)

Not specified

4 (23.5%)

*Percentages for dependent variables do not add up to 100% because some studies
include several dependent variables.

Another interesting point worth men-

tioning is with regards to the relationship
between the published and unpublished
studies and the quasi- and non-experimental
study designs. It was found that of the 12
published articles, eight (67.7 per cent) used
quasi-experimental designs, while only four
(33.3 per cent) used non-experimental
designs, compared to the five unpublished
articles which only had of two (40.0 per
cent) studies using quasi-experimental
designs, and three (60.0 per cent) studies
using non-experimental designs.

Overall
Initially, an analysis was undertaken to deter-
mine if the sample of effect sizes corrected to
eliminate multiple endpoints (N=80) – and
used in all analyses – yielded results different
from the original total population of effect
sizes (N=180). As Table 2 shows, the overall
analysis of the 180 effect sizes revealed a sig-
nificant moderate positive effect; Hedges
g=.451, p<.05. After the adjustment for multi-
ple endpoints, the overall effect size (Hedges
g) was .427, p<.05; the difference between
these was not statistically significant (p>.05).
Thus, the overall effectiveness of team build-
ing interventions in sport (i.e. for all
dependent variables combined) was found
to be positive and moderate in size.

Characteristics of the study

Experimental design. The first question of
interest with regards to the characteristics of
the study was whether the type of design
used yielded different conclusions with
regard to the effectiveness of team building
in sport. As Table 2 shows, results from
studies using a quasi-experimental design
and those using a non-experimental design
produced similar (p>.05) moderate positive
effects (Hedges g=.408 and .474 respec-
tively), as well as Fail Safe N’s of 121, and
144.

Source. Published (Hedges g=.385) and non-
published (Hedges g=.539) studies did show
a seemingly different picture about the effec-

tiveness of team building interventions in
sport and had slightly different Fail Safe N’s
of 114 and 166 respectively. However, these
differences were not statistically significant
(see Table 2).

Characteristics of the intervention

Intervention type. As the results in Table 2
show, goal setting and adventure programme
techniques both produced significantly
(p<.05) positive results (Hedges g=.714 and
.417 respectively). Conversely, team building
interventions concerned with interpersonal
relations (Hedges g=.486) and a variety of
outcomes (i.e. combination interventions,
Hedges g=.161) failed to produce statistically
significant (p>.05) effects.

An analysis of the differences among

intervention types was statistically significant
(L(3)=14.73, p<.01). Bonferroni post-hoc
analyses showed one significant difference –
goal setting interventions were significantly
superior to interventions that focused on a
variety of factors (p<.001). The Fail Safe N’s
were as follows, 226 for goal setting, 38 for
combination, 148 for interpersonal rela-
tions, and 143 for adventure programmes.

Mode of intervention delivery. As indicated
above, team building interventions in sport
typically use one of two approaches for deliv-
ery of the programme (i.e. direct vs. indi-
rect). As Table 2 shows, both methods are
equally effective, producing significant
(p<.05) moderate positive effects (direct,
Hedges g=.446; indirect, Hedges g=.414) that
did not differ from one another statistically
(p>.05). Their Fail Safe N statistics were also
very similar at 135 and 124 respectively.

Intervention length. As Table 2 shows, inter-
vention length does have an influence on
outcome effectiveness. Interventions lasting
less than two weeks had a non significant
impact (Hedges g=.106, p>.05). On the other
hand, interventions ranging in duration
from two to 20 weeks and 20 weeks or longer
both resulted in significant moderate effects
(Hedges g=.499 and .564 respectively, p<.05),

Sport & Exercise Psychology Review, Vol. 5 No. 2

11

Team building interventions in sport: A meta-analysis

12

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Table 2: The effectiveness of team building in sport.

Fail

Effect

Number of

Significance

Safe

Factor

L-statistic

size

SE

effect sizes

(p-value)

N

Overall

L(1)=.21

Original sample

p>.05, ns

.451

.056

180

0.000

136

Final sample

.427

.072

80

0.000

128

Design

L(1)=.00

Quasi-experimental

p>.05, ns

.408

.096

49

0.000

121

Non-experimental

.474

.091

31

0.000

144

Source

L(1)=.12

Published

p>.05, ns

.385

.091

59

0.000

114

Unpublished

.539

.104

21

0.000

166

Type of intervention

L(3)=14.73

Goal setting

p<.01

.714

.103

22

0.000

226

Combination

.161

.140

29

ns

38

Interpersonal relations

.486

.313

11

ns

148

Adventure programmes

.471

.110

18

0.000

143

Delivery

L(1)=.93

Direct

p>.05, ns

.446

.086

39

0.000

135

Indirect

.414

.125

41

0.001

124

Length of intervention

L(2)=6.90

Less than 2 weeks

p<.05

.106

.157

19

ns

19

2 to 20 weeks

.499

.115

26

0.000

153

20 weeks and above

.564

.103

35

0.000

175

Gender

L(2)=1.19

Male

p>.05, ns

.334

.132

28

0.000

97

Female

.458

.090

47

0.000

139

Mixed

.712

.224

5

0.002

225

Sport type

L(2)=13.85

Interactive

p<.01

.159

.092

34

ns

37

Individual

.673

.105

41

0.000

212

Combination

.712

.224

5

0.002

225

Skill level

L(2)=3.53

High School

p>.05, ns

.240

.135

4

ns

65

Intercollegiate

.482

.088

66

0.000

147

Post-intercollegiate

.218

.138

10

ns

57

Dependent variables

L(5)=19.37

Social cohesion

p<.01

.214

.069

21

0.002

56

Task cohesion

.263

.186

22

ns

72

Performance

.712

.132

14

0.000

225

Enhanced cognitions

.799

.094

14

0.000

255

Roles

.789

.537

4

ns

251

Anxiety

–.165

.393

5

ns

39

Note: ns=not significant

as well as comparable Fail Safe N statistics
(135 and 175 respectively). The difference
among these effect sizes was significant
(L(2)=6.90, p<.05); Bonferroni post-hoc
analyses revealed that interventions lasting
less than two weeks were significantly infe-
rior (p<.05) to interventions that were
longer than 20 weeks in duration.

Characteristics of the participants

Gender. As Table 2 shows, team building
interventions had a significant (p<.05) bene-
ficial impact for male teams (Hedges g=.334)
(Fail Safe N, 97), female teams (Hedges
g=.458) (Fail Safe N, 139), and in samples
composed of both female and male teams
(Hedges g=.712) (Fail Safe N, 225). An analy-
sis of the differences among these three cat-
egories of participants revealed no
significant differences (p>.05).

Type of sport. Team building had less impact
on sport teams categorised as interactive
(Hedges g=.525, p>.05) than for teams cate-
gorised as individual (Hedges g=.673, p<.05)
and combination (Hedges g=.712, p<.05).
Significant differences were found among
the three categories (L(2)=13.85, p<.01);
Bonferroni post-hoc analysis showed that
team building interventions delivered to
individual sport teams were more beneficial
than those delivered to interactive sport
teams (p<.05). The Fail Safe N statistics cal-
culated were 37, 212, and 225 respectively.

Skill level. The category of teams that showed
significant team building effects was at the
intercollegiate level (Hedges g=.482, p<.05);
the effect sizes were non-significant (p>.05)
for both the high school (Hedges g=.240)
and post-collegiate levels (Hedges g=.218).
Possibly because of the small number of
effect sizes available in both the high school
and post-collegiate conditions and/or the
relatively large variability in the results, no
significant differences were found between
these three categories (L(2)=3.53, p>.05).
The Fail Safe N statistics calculated were 65,
147, and 57 respectively.

Dependent variables

As Table 2 shows, team building interven-
tions had a different effect on different cate-
gories of outcomes. Also, considerable
variability is present within some categories.
Specifically, team building had no significant
effect on task cohesion (Hedges g=.263,
p>.05), roles (Hedges g=.789, p>.05) or anxi-
ety (Hedges g=–.165, p>.05). There is a small
positive effect for social cohesion (Hedges
g=.214, p<.05) and large positive effects for
both performance (Hedges g=.712, p<.05)
and enhanced cognitions (Hedges g=.789,
p<.05).

Further analyses showed that significant

differences were present among the types of
dependent variables (L(5)=19.37, p<.01).
Bonferroni post-hoc tests revealed that team
building had significantly less influence on
social cohesion than both performance
(p<.05) and enhanced cognitions (p<.01).
A final calculation for the Fail Safe N statis-
tics showed numbers of 56 for social cohe-
sion, 72 for task cohesion, 225 for
performance, 255 for enhanced cognitions,
251 for roles, and finally, 39 for anxiety.

Discussion

The primary purpose of the study was to con-
duct a meta-analysis to examine the effective-
ness of team building interventions in sport.
A secondary purpose was to examine the
influence of a number of different modera-
tor variables. The specific moderator vari-
ables examined fell under four general
categories: characteristics of the study, char-
acteristics of the team building intervention,
characteristics of the participants, and
dependent variables. A number of the find-
ings pertaining to each of these warrants dis-
cussion. The first pertains to the surprising
results associated with study design (i.e. no
differences were observed between studies
employing quasi-experimental and studies
employing non-experimental designs).
Fraker and Maynard (1987) have pointed out
that non-experimental conditions are not as
reliable as conditions including a control
group. Perhaps one of the reasons for the

Sport & Exercise Psychology Review, Vol. 5 No. 2

13

Team building interventions in sport: A meta-analysis

lack of difference observed in the present
results is – as DeMeuse and Liebowitz (1981)
discussed – that the control group athletes
felt compelled to outperform the experimen-
tal group athletes. This would decrease the
apparent effect of the intervention.

The second set of findings that should be

highlighted is associated with the sources of
data. In the present study, it was observed
that published and unpublished articles pro-
duced identical positive moderate effects.
Rosenthal (1979) directed attention toward
the possibility that journals might publish
only statistically significant findings. This
does not appear to be an issue in the area of
team building in sport. Therefore, confi-
dence in the pattern of findings found in
published (and unpublished) reports is justi-
fied.

The third set of results that warrants dis-

cussion is associated with the types of interven-
tions typically used in sport. The two most
frequently used interventions in the literature
are goal setting (N=22 effect sizes) and
omnibus programmes (N=29 effect sizes).
Goal setting programmes were associated with
significant large effects that were also signifi-
cantly superior to the small non-significant
effects emanating from omnibus pro-
grammes. The effectiveness of goal setting
should not be surprising as the results from
meta-analyses on individual (Kyllo & Landers,
1995) and group goal setting (O’Leary-et al.,
1994) have shown to be effective. Also, as
Locke and Latham have pointed out goal set-
ting has been shown to be effective ‘on well
over 100 different tasks involving more than
40,000 participants in at least eight countries
working in laboratory, simulation, and field
settings’ (Locke & Latham, 2002, p.714).

One possible reason for the difference

between goal setting and omnibus interven-
tions may reside in the attention required on
the part of participants in the two types of
programmes. With goal setting programmes,
a specific (and generally limited) set of
behaviours is targeted; thus, the focus of ath-
letes’ attention and actions is specific, nar-
row, and directed (e.g. Locke & Latham,

2006). With omnibus programmes, on the
other hand, a general (and usually broad)
set of variables is targeted (see Prapavessis et
al., 1996, for an example of such a pro-
gramme); thus, the focus of the athletes is
more diffuse. A point for coaches and practi-
tioners to consider is that if fewer interven-
tions are introduced, athletes may be better
able to focus their attention, and, in turn,
more positive outcomes may result.

Another finding that warrants attention

also has implications for coaches and practi-
tioners. It was found that direct and indirect
methods of delivery did not differ in effective-
ness. Both produced positive moderate
effects. Consequently, coaches have the free-
dom to choose which method they feel
would be more effective for their team. For
example, sometimes it may be beneficial to
introduce a new person (i.e. sport consult-
ant) to keep things fresh and interesting,
whereas in other situations, it may negatively
impact the cohesion of a group so the coach
ought to manage the intervention.

A fifth set of findings that warrants atten-

tion pertains to the ineffectiveness of inter-
ventions of short duration (i.e. less than two
weeks). These findings lend support to sug-
gestions advanced by reviewers of the team
building literature in industrial psychology
(DeMeuse & Liebowits, 1981) and sport psy-
chology (Brawley & Paskevich, 1997). Specif-
ically, DeMeuse and Liebowits (1981)
proposed that interventions in business and
industry that last less than one year are not
effective. Similarly, Brawley and Paskevich
(1997) proposed that the length of a sport
team intervention should be no less than
one season in length. The results of the pres-
ent study showed that as the length of the
intervention increased (from less than two
weeks to greater than 20 weeks), effective-
ness improved. It seems probable that it
takes time for an intervention to be properly
introduced to a team, for athletes to gain
trust in the programme, and for behaviour
change to occur. Interventions of minimal
duration seemingly do not afford the neces-
sary time.

14

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

The results for gender showed that team

building is equally effective for teams com-
posed of females and males. In sport, team
building might take different forms for male
and females; the present results show, how-
ever, that both profit from this type of inter-
vention.

The type of sport, however, is a moderator

variable in the team building-outcome rela-
tionship; team building is significantly more
beneficial in individual sports than in team
sports. This result supports the suggestion
advanced by Carron et al. (2002), that team-
building interventions might have a greater
impact on both team cohesion and team per-
formance in individual sport contexts.

A ceiling effect may be in operation.

Although interactive teams do profit from
team building, in many instances they are
already cohesive so further improvement
may be minimal. Individual teams, however,
may not be as cohesive initially and as a con-
sequence, a much more significant change
in cohesion is possible.

The final set of results that should be dis-

cussed pertains to the dependent variables
analysed. It is interesting to note that the
most frequently examined variable, cohesion
(representing 53 per cent of the effect sizes
examined), was among the least influenced
by team building. One possible reason for
this could stem from the measures that were
included in the task and social cohesion
(moderator variable) categories – specifi-
cally, some studies assessed cohesion using
the psychometrically sound Group Environ-
ment Questionnaire (e.g. Rainey & Schwe-
ickert, 1988; McClure & Foster, 1991; Stevens
& Bloom, 2003), whereas other studies used
more general (in terms of cohesion) meas-
ures such as attraction to the group, unity of
purpose, quality of teamwork, and athletic
communication (e.g. Hoigaard et al., 2006;
Morgan, 2006). Regardless of the meas-
ure(s) used, this finding is intriguing
because although people continue to use
team building in hopes of increasing cohe-
sion, it may not have the desired impact.

Another point of interest was that team

building interventions targeting enhanced
cognitions also had a large positive effect,
which further strengthens the belief that psy-
chological, physical, and physiological skills
do respond to practice (e.g. Arvinen-Barrow
et al., 2008; Tremayne & Ballinger, 2008;
Rascle et al., 2008). These results are very
promising for sport psychologists and practi-
tioners attempting to provide effective inter-
ventions.

The present meta-analysis is not without

its limitations. First, the majority of studies
(82.4 per cent; N=14) targeted intercolle-
giate teams and athletes; only three studies
(Martin & Davids, 1995; Prapavessis et al.,
1996; Senecal et al., 2007) used other types
of samples. The reason for this may be that
most research is done in the university set-
ting and there are close ties and relation-
ships with the university sport coaches. This
is very positive in the sense that the research
benefits from the availability of elite compet-
itive athletes from a rich pool of sports.
Future research, however, should attempt to
study different teams at different skill levels
to determine if the findings generalise across
all sport.

Another limitation is common to most

meta-analyses. The studies incorporated
varied in factors such as length of the team
building intervention, the type of sport, the
different outcomes assessed, gender of par-
ticipants, and so on. These differences in the
studies represent one of the classic criticisms
of meta-analyses; basically, that they are a
‘mixing of apples and oranges’ (Thomas &
French, 1986).

The results of the present study provide

implications for anyone attempting to con-
duct a team building intervention in sport.
Interventions are costly and time consum-
ing, therefore, the desire for the sport con-
sultant or coach should be to give the
intervention the best possibility of success.
Specifically, it appears that the most success-
ful interventions should: (a) use either goal
setting or adventure programme techniques
(note that although the adventure pro-

Sport & Exercise Psychology Review, Vol. 5 No. 2

15

Team building interventions in sport: A meta-analysis

grammes were effective, they were all of
short duration); (b) be at least two weeks in
duration, and preferably over 20 weeks
(depending on the season length of the
sport in question); (c) target individual
teams (if cohesion is the overall goal of the
intervention); and (d) aim to increase per-
formance or improve cognitions.

The authors

Luc J. Martin ([e-mail]),
Albert V. Carron and
Shauna M. Burke
all from the School of Kinesiology,
University of Western Ontario,
Canada.

16

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Anderson, C.D., Crowell, C.R., Doman, M. &

Howard, G.S. (1988)*. Performance posting,
goal setting, and activity-contingent praise as
applied to a university hockey team. Journal of
Applied Psychology, 73, 87–95.

Arvinen-Barrow, M., Weigand, D.A., Hemmings, B. &

Walley, M. (2008). The use of imagery across
competitive levels and time of season: A cross-
sectional study among synchronised skaters in
Finland. European Journal of Sport Science, 8,
135–142.

Bangert-Drowns, R.L. (1986). Review of development

in meta-analytic method. Psychological Bulletin, 99,
388–399.

Baumeister, R.F. & Leary, M.R. (1995). The need to

belong: Desire for interpersonal attachments as a
fundamental human motivation. Psychological
Bulletin, 117, 497–529.

Beer, M. (1980). Organisational change and development:

A systems review. Glenview, IL: Scott, Foresman.

Blake, R. & Mouton, J. (1964). The managerial grid.

Houston: Gulf Publishing.

Bloom, G.A. & Stevens, D.E. (2002)*. A team

building mental skills training programme with
an intercollegiate equestrian team: Athletic
Insight. The Online Journal of Sport Psychology, 4.
www.athleticinsight.com/Vol4Issl/
Applied_Issue.htm

Brawley, L.R. & Paskevich, D.M. (1997). Conducting

team building research in context of sport and
exercise. Journal of Applied Sport Psychology, 9,
11–40.

Buller, P.F. & Bell, C.H. (1986). Effects of team

building and goal setting on productivity: A field
experiment. Academy of Management Journal, 29,
305–328.

Burke, S.M., Carron, A.V., Eys, M.A., Ntoumanis, N. &

Estabrooks, P.A. (2006). Group versus individual
approach? A meta-analysis of the effectiveness of
interventions to promote physical activity.
Sport and Exercise Psychology Review, 2, 19–35.

Burton, D. (1989)*. Winning isn’t everything:

Examining the impact of performance goals on
collegiate swimmers’ cognitions and
performance. The Sport Psychologist, 3, 105–132.

Carron, A.V., Colman, M.M., Wheeler, J. & Stevens,

D. (2002). Cohesion and performance in sport:
A meta-analysis. Journal of Sport and Exercise
Psychology, 24, 168–188.

Carron, A.V. & Hausenblas, H.A. (1998). Group

dynamics in sport (2nd ed.). Morgantown, WV:
Fitness Information Technology, Inc.

Carron, A.V., Hausenblas, H.A. & Eys, M.A. (2005).

Group dynamics in sport (3rd ed.). Morgantown,
WV: Fitness Information Technology, Inc.

Carron, A.V. & Spink, K.S. (1993)*. Team building in

an exercise settings. The Sport Psychologist, 7, 8–18.

Chelladurai, P. (1978). A contingency model of leadership

in athletics. Unpublished doctoral dissertation.
Waterloo: University of Waterloo, Canada.

Cohen, J. (1988). Statistical power analysis for the

behavioural sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.

Cogan, K.D. & Petrie, T.A. (1995)*. Sport

consultation: An evaluation of a season-long
intervention with female collegiate gymnasts.
The Sport Psychologist, 9, 282–296.

DeMeuse, K.P. & Liebowitz, S.J. (1981). An empirical

analysis of team building research. Group and
Organization Studies, 6, 357–378.

Duda, J.L. (1987). Toward a developmental theory of

children’s motivation in sport. Journal of Sport and
Exercise Psychology in Sport, 9, 130–145.

Dunn, J.G. & Holt, L.H. (2004). A qualitative

investigation of a personal-disclosure mutual-
sharing team building activity. The Sport
Psychologist, 18, 363–380.

Eagly, A.H. & Karau, S.J. (1991). Gender and the

emergence of leaders: A meta-analysis. Journal of
Personality and Social Psychology, 60, 685–710.

Eden, D. (1986). Team development: Quasi-

experimental confirmation among combat
companies. Group and Organizational Studies, 11,
133–146.

Estabrooks, P.A. & Carron, A.V. (1999). Group

cohesion in older adult exercisers: Prediction
and intervention effects. Journal of Behavioral
Medicine, 22, 575–588.

Field, A.P. (2001). Meta-analysis of correlation

coefficients: A Monte Carlo comparison of fixed-
and random-effects methods. Psychological
Methods, 6, 161–180.

Fraker, T. & Maynard, R. (1987). The adequacy of

comparison group designs for evaluation of
employment-related programmes. The Journal of
Human Resources, 22, 194–227.

Gill, D.L. (1986). Competitiveness among females

and males in physical activity classes. Sex Roles, 15,
233–247.

Gleser, L.J. & Olkin, I. (1994). Stochastically

dependent effect sizes. In H. Cooper & L.V.
Hedges (Eds.), The handbook of research synthesis
(pp.339–357). New York: Russell Sage
Foundation.

Hedges, L.V. (1981). Distribution theory for Glass’s

estimator of effect size and related estimators.
Journal of Educational Statistics, 6, 107-128.

Hedges, L.V. (1982). Estimation of effect size from a

series of independent experiments. Psychological
Bulletin, 92, 490–499.

Sport & Exercise Psychology Review, Vol. 5 No. 2

17

Team building interventions in sport: A meta-analysis

References

Note: References highlighted by an asterisk (*) were included in the meta-analysis.

18

Sport & Exercise Psychology Review, Vol. 5 No. 2

Luc J. Martin, Albert V. Carron & Shauna M. Burke

Hedges, L.V. & Olkin, I. (1985). Statistical methods for

meta-analysis. New York: Academic Press.

Hoigaard, R., Tofteland, I. & Ommundsen, Y. (2006)*.

The effect of team cohesion on social loafing in
relay teams. International Journal of Applied Sports
Sciences, 18, 59–73.

Holt, L.H. & Dunn, J.G. (2006). Guidelines for

delivering personal-disclosures mutual-sharing
team building interventions. The Sport
Psychologist, 20, 348–367.

Hunter, J.E. & Schmidt, F.L. (2000). Fixed effects vs.

random effects meta-analysis models:
Implications for cumulative research knowledge
in psychology. International Journal of Selection and
Assessment, 8, 275–292.

Kilty, K.M. (2000)*. A study of cohesion in women’s sports

teams using adventure programming. Unpublished
doctoral dissertation, Boston University, Boston,
United States.

Kyllo, L.B. & Landers, D.M. (1995). Goal setting in

sport and exercise: A research synthesis to
resolve the controversy. Journal of Sport and
Exercise Psychology, 17, 117–137.

Locke, E.A. & Latham, G.P. (2002). Building a

practically useful theory of goal setting and task
motivation: A 35-year odyssey. American
Psychologist, 57, 705–717.

Locke, E.A. & Latham, G.P. (2006). New directions in

goal-setting theory. Current Directions in
Psychological Science, 15, 265–268.

Martens, R. & Burton, D. (1982). Psychological skills

training for athletes. Unpublished manuscript.
University of Illinois.

Martin, R. & Davids, K. (1995)*. The effects of group

development techniques on a professional
athletic team. The Journal of Social Psychology, 135,
533–536.

McClure, B.A. & Foster, C.D. (1991)*. Group work as

a method of promoting cohesiveness within a
women’s gymnastics team. Perceptual and Motor
Skills, 73, 307–313.

Morgan, T.K. (2006)*. A season-long mental skills

training programme for collegiate volleyball players.
Unpublished doctoral dissertation, University of
Tennessee, Knoxville, United States.

Murphy, J.M. (2001)*. Effect of a one-time team building

exercise on team cohesion when working with a NCAA
Division 1 women’s basketball team. Unpublished
masters of science thesis, North Carolina State
University, North Carolina, United States.

Neuman, G.A., Edwards, J.E. & Raju, N.S. (1989).

Organisational development interventions:
A meta-analysis of their effects on satisfaction
and other attitudes. Personal Psychology, 42,
461–483.

O’Leary-Kelly, A.M., Martocchio, J.J. & Frink, D.D.

(1994). A review of the influence of group goals
on group performance. Academy of Management
Journal, 37, 1285–1301.

Pangos, W. (1986)*. Goal setting and its effects on

basketball performance. MA Thesis, University of
Western Ontario, London, Ontario.

Prapavessis, H., Carron, A.V. & Spink, K.S. (1996)*.

Team building in sport. International Journal of
Sport Psychology, 27, 269–285.

Rainey, D.W. & Schweickert, G.J. (1988)*. An

exploratory study of team cohesion before and
after a Spring trip. The Sport Psychologist, 2,
314–317.

Rascle, O., Le Foll, D. & Higgins, N.C. (2008).

Attributional retraining alters novice golfers’ free
practice behaviour. Journal of Applied Sport
Psychology, 20, 157–164.

Roque da Costa Rolo, C.M. (2004)*. An intervention

for fostering hope, athletic and academic performance
in university student-athletes. Unpublished doctoral
dissertation, University of North Carolina,
Greensboro, United States.

Rosenthal, R. (1966). Experimenter effects in behavioural

research. New York: Appleton-Century-Crofts.

Rosenthal, R. (1979). The ‘file drawer problem’ and

tolerance for null results. Psychological Bulletin,
86, 638–641.

Schurr, K.T., Ashley, M.A. & Joy, K.L. (1997).

A multivariate analysis of male athlete
characteristics: Sport type and success.
Multivariate Experimental Clinical Research, 3,
53–68.

Senecal, J., Loughead, T.M. & Bloom, G.A. (2007)*.

A season-long team building intervention
programme: Examining the effect of team goal
setting on cohesion. Journal of Sport and Exercise
Psychology, 30, 186–199.

Stevens, D.E. & Bloom, G.A. (2003)*. The effect of

team building on cohesion. Avante, 9, 43–54.

Thomas, J.R. & French, K.E. (1986). The use of meta-

analysis in exercise and sport: A tutorial. Research
Quarterly for Exercise and Sport, 57, 196–204.

Thomas, J.R., Nelson, J.K. & Thomas, K.T. (1999).

A generalised rank-order method for non-
parametric analysis of data from exercise science:
A tutorial. Research Quarterly for Exercise and Sport,
70, 11–23.

Tremayne, P. & Ballinger, D.A. (2008). Performance

enhancement for ballroom dancers:
Psychological perspectives. The Sport Psychologist,
22, 90–108.