Is team confidence the key to success? The reciprocal relation

Is team confidence the key to success? The reciprocal relation

Running head: RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 1 2 3 4 5 6 7 8 Is team confidence the key to success? The reciprocal relation between...

309KB Sizes 0 Downloads 0 Views

Recommend Documents

Competitiveness is the Key to Success
as Minsk Automobile Plant (MAZ trademark), Minsk Engine Works. (MMZ trademark), Vitebsk Plant of Electrical Measuring In

key to Club success
A Well-Run Club is The Key to a Successful Club. Herb Cartmell, AU Constitution and Bylaws Advisor. We all hate to have

key to success - EuroChem
EuroChem Moscow Office. A KGOK branch in Perm Territory. EuroChem-VolgaKaliy LLC ..... University) and MGUIE (Moscow Sta

dedication to customers is key to kehoe success - Kehoe Brothers
Kehoe. Massey Ferguson. Denis explained: “The Massey. Ferguson tractor franchise has been ... were (from left): Andrew

Asia on the Buyside: The Key to Success - Kroll
asia on the Buyside: THE KEY TO SuccESS. Foreword. 02. Executive summary. 03. Survey findings. 05. Asia-Pacific's M&A la

Resiliency: A Key to Success
Mar 31, 2017 - needs of our clients to facilitate the building ... of capability – 5 building blocks of resilience. â–

key to our success - EMSB
Apr 16, 2013 - Lemieux, Linda. Lepine, Renée. Leprohon, Vicky. Ligris, Bessie. Loupessis, Katerina. Loureiro, Ana Paula.

Ingram Micro SMB 500 :: Location Is One Key to Success
In Lakewood, N.J., a town once referred to as “peaceful and pleasure-loving” by ... Silicon Valley continues to thri

The Up Side Of Down: Why Failing Well Is The Key To Success By
Down: Why Failing Well Is the Key to Success is true to its title, flipping an entrenched view of success on its ear. Au

The Up Side Of Down: Why Failing Well Is The Key To Success By
If you are searched for a book The Up Side of Down: Why Failing Well Is the Key to Success by Megan. McArdle in pdf form

Running head: RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 1 2 3 4 5 6 7 8

Is team confidence the key to success? The reciprocal relation between collective efficacy,

9

team outcome confidence, and perceptions of team performance during soccer games.

10 11

Katrien Fransen1, Steven Decroos1, Norbert Vanbeselaere2, Gert Vande Broek1,

12

Bert De Cuyper1, Jari Vanroy1, & Filip Boen1 1

13 14

2

Department of Kinesiology, KU Leuven, Leuven, Belgium

Center for Social and Cultural Psychology, KU Leuven, Leuven, Belgium

15 16 17

Journal of Sports Sciences, Manuscript in press

18 19 20 21 22

Acknowledgements

23

This research was supported by a PhD Fellowship (Aspirant) of the Research Foundation

24

Flanders (FWO), awarded to Katrien Fransen. We are grateful to Lode Wuyts for his

25

assistance during the data collection.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 26 27

2

Abstract The present manuscript extends previous research on the reciprocal relation between

28

team confidence and perceived team performance in two ways. First, we distinguished

29

between two types of team confidence; process-oriented collective efficacy and outcome-

30

oriented team outcome confidence. Second, we assessed both types not only before and after

31

the game, but for the first time also during half-time, thereby providing deeper insight into

32

their dynamic relation with perceived team performance. Two field studies were conducted,

33

each with 10 male soccer teams (N = 134 in Study 1; N = 125 in Study 2). Our findings

34

provide partial support for the reciprocal relation between players‟ team confidence (both

35

collective efficacy and team outcome confidence) and players‟ perceptions of the team‟s

36

performance. Although both types of players‟ team confidence before the game were not

37

significantly related to perceived team performance in the first half, players‟ team confidence

38

during half-time was positively related to perceived team performance in the second half.

39

Additionally, our findings consistently demonstrated a relation between perceived team

40

performance and players‟ subsequent team confidence. Considering that team confidence is a

41

dynamical process, which can be affected by coaches and players, our findings open new

42

avenues to optimize team performance.

43 44

Keywords: winning confidence, in-game measurements, continuous measurements, team dynamics, sport psychology

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 45

3

Introduction

46

Coaches, players and other team sport enthusiasts often mention team confidence as a

47

key to success; “What you believe, you can achieve” (Quinn, 2012, p. 90). Research findings

48

confirmed these on-field perceptions by demonstrating that athletes who were more confident

49

in their team‟s abilities exerted more effort (Greenlees, Graydon, & Maynard, 1999), set more

50

challenging goals (Silver & Bufanio, 1996), were more resilient when facing adversities

51

(Morgan, Fletcher, & Sarkar, 2013), and ultimately performed better (Stajkovic, Lee, &

52

Nyberg, 2009).

53

Although these findings stress the importance of team confidence, the existing

54

literature is characterized by inconsistencies in the way in which the construct of team

55

confidence has been conceptualized, operationalized, and measured (Shearer, Holmes, &

56

Mellalieu, 2009). Overall, two distinct types of team confidence can be identified (Collins &

57

Parker, 2010; Fransen, Kleinert, Dithurbide, Vanbeselaere, & Boen, 2014). The first type has

58

been termed collective efficacy and was originally defined by Bandura (1997, p. 477) as “a

59

group‟s shared belief in its conjoint capability to organize and execute the courses of action

60

required to produce given levels of attainment”. In other words, collective efficacy comprises

61

athletes‟ confidence in the process of their own team, rather than comparing their own

62

abilities with those of the opposing team. Consequently, collective efficacy has been

63

measured as athletes‟ confidence in the skills of their team required to accomplish a certain

64

task (e.g., “I believe that my team will demonstrate a strong work ethic during this game”).

65

In contrast, the second type of team confidence focuses on outperforming the

66

opponent and comprises athletes‟ confidence in their team‟s abilities to obtain a certain

67

outcome (e.g., “I believe that my team will win this game”). Collins and Parker (2010) termed

68

this construct „team outcome efficacy‟. In sports, this outcome-oriented confidence in winning

69

or performing better than the opponent has been termed „competitive efficacy‟ or

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

4

70

„comparative efficacy‟ (Myers & Feltz, 2007). However, this outcome-oriented measure does

71

not capture the process-oriented nature of collective efficacy as described by Bandura (1997).

72

As such, an „efficacy‟ label seems inappropriate. Moreover, several authors emphasized the

73

difference between the confidence in outperforming the opponent (i.e., performance

74

judgments) and outcome expectations (Myers & Feltz, 2007; Myers, Paiement, & Feltz,

75

2007). Bandura (1997, pp. 22-23) noted that “an outcome is the consequence of a

76

performance, not the performance itself.” Performance accomplishments can take the form of

77

letter grades in academia or a final game score in sports. A trophy, praise from the coach, or

78

self-satisfaction are examples of outcomes that might ensue from a performance

79

accomplishment (Myers & Feltz, 2007). Given the conceptual differences between efficacy

80

beliefs and outcome expectations, the outcome-oriented measure of team confidence has

81

recently been labeled „team outcome confidence‟ (Fransen, Kleinert, et al., 2014). We adopt

82

this recent conceptualization in the current research and distinguish between „process-oriented

83

collective efficacy‟ on the one hand and „outcome-oriented team outcome confidence‟ on the

84

other hand.

85

Although a number of studies have confirmed the reciprocal relation between team

86

confidence and performance (for a meta-analysis see Stajkovic et al., 2009), the difference

87

between process- and outcome-oriented team confidence has been disregarded. Moreover, a

88

number of studies used the outcome-oriented measurement to allegedly assess collective

89

efficacy (e.g., Chen et al., 2002; Fransen et al., 2012; Spink, 1990; Tasa, Taggar, & Seijts,

90

2007; Vargas-Tonsing & Bartholomew, 2006). Therefore, the present manuscript will go one

91

step further by examining the reciprocal relation between performance and both collective

92

efficacy and team outcome confidence.

93 94

In order to ground our hypotheses on the existing literature, previous studies had to be interpreted with regard to the measurements they used to assess the team

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

5

95

confidenceperformance relation. Based on the distinction described earlier, we classified

96

previous studies as targeting either collective efficacy or team outcome confidence. First, with

97

regard to collective efficacy, the literature review revealed inconsistent results regarding its

98

relation with team performance. Bandura (1997, p. 470) stated: “the higher the sense of

99

collective efficacy, the better the team‟s performance”. A meta-analytic review including 96

100

studies, confirmed this statement and revealed that collective efficacy is significantly related

101

to group performance (Stajkovic et al., 2009). In line with these findings, Keshtan,

102

Ramzaninezhad, Kordshooli, and Panahi (2010) demonstrated that professional volleyball

103

teams with high levels of collective efficacy were positioned higher in the ranking than

104

professional teams with low levels of collective efficacy. In contrast, a study with university

105

basketball teams revealed no significant relation between a team‟s collective efficacy and the

106

team‟s performance, measured by shooting percentage and difference in rebounds taken

107

(MacLean & Sullivan, 2003). Likewise, Chen et al. (2002) revealed that in more recreational

108

basketball teams players‟ collective efficacy did not predict the team‟s performance, assessed

109

by the season winning percentage and the point difference.

110

Second, with regard to team outcome confidence, the literature consistently revealed a

111

positive relation with performance. In the experiment of Stanimirovic and Hanrahan (2004),

112

teams of secondary school students were assigned to either a repeated success or repeated

113

failure condition. Success and failure were manipulated by having participants compete

114

against a respectively lower or higher score of an imaginary opponent. The results

115

demonstrated the positive impact of performance on team outcome confidence; teams in the

116

repeated success condition reported higher confidence in winning the game than teams

117

competing in the repeated failure condition. On the other hand, two laboratory studies

118

revealed that the reversed causal direction also holds since they observed that teams with a

119

higher team outcome confidence performed better than teams who lost confidence in their

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

6

120

winning chances (Chen et al., 2002; Hodges & Carron, 1992). Additionally, field studies in

121

intercollegiate ice hockey teams delivered further support for the reciprocal relation between

122

team outcome confidence and team performance, measured by official game statistics (Feltz

123

& Lirgg, 1998; Myers, Paiement, & Feltz, 2004).

124

Besides the inconsistencies in how team confidence has been assessed, another

125

shortcoming in the current literature relates to the timing of the measurement. Team

126

confidence has been conceptualized as a dynamic construct, rather than as a trait-like

127

characteristic showing strong cross-temporal stability (Myers & Feltz, 2007). In other words,

128

players‟ confidence in their team‟s abilities may change in the course of the game, and these

129

changes may impact on winning or losing. Therefore, Bandura (1997, p. 67) stated that the

130

relation between team confidence and performance is revealed most accurately when both

131

constructs are measured in close temporal proximity.

132

Myers, et al. (2007) tested the importance of this temporal proximity by examining the

133

relation between team confidence, measured before the game, and three cumulative

134

performance intervals within ice hockey games. Their results revealed that team confidence

135

before the game was a significant predictor of team performance at each of the three

136

performance intervals. However, the magnitude of this relationship did not change

137

significantly as the temporal proximity between team confidence and performance decreased.

138

It should be noted though that team confidence was only measured once within the 24 hours

139

before the game. In the time span between the measurement of team confidence and the

140

team‟s performance, intervening experiences may have impacted on the players‟ confidence

141

(e.g., a coach‟s motivational speech or the playing level of the team). As a consequence, it has

142

been suggested that the best way to minimize this problem is to measure players‟ team

143

confidence during performance (Myers & Feltz, 2007).

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 144

7

Despite these guidelines and disregarding the dynamic nature of team confidence, the

145

concept of team confidence has traditionally been measured as a trait concept or, at best,

146

before or after a game, but not during a game. The only exception is a study by Edmonds,

147

Tenenbaum, Kamata, and Johnson (2009) in which team confidence was measured at three

148

time points during an adventure race. Their results partially supported the dynamic view on

149

the team confidenceperformance relation; the higher athletes‟ confidence before each

150

discipline, the better they performed at it. However, because the race consisted of five

151

different disciplines (i.e., trekking, canoeing, mountain biking, climbing, and orienteering),

152

the effects of a previous performance on the team‟s confidence in successfully accomplishing

153

a subsequent task were very small. This variety in the disciplines involved in the adventure

154

race makes it dangerous to generalize the results to sport teams in which players perform a

155

similar task during the entire game (e.g., soccer).

156

In line with previous recommendations (Bandura, 1997; Myers & Feltz, 2007), the

157

present research took a first step toward a more dynamic in-game measurement of players‟

158

team confidence. Therefore, we measured players‟ team confidence at different time points,

159

but, in contrast to Edmonds et al. (2009), within the same task (i.e., a soccer game). In Study

160

1, both types of team confidence (i.e., collective efficacy and team outcome confidence) were

161

measured before the game and at the start and the end of the half-time break. In this way, we

162

tried to account for the speech of the coach during half-time, because it has already been

163

argued that verbal persuasion is one of the most effective methods for coaches to build team

164

confidence (Fransen et al., 2012; Vargas-Tonsing & Bartholomew, 2006; Vargas-Tonsing,

165

Myers, & Feltz, 2004). In Study 2, measurements of team confidence after the game were

166

added, thereby aiming at a deeper insight in the dynamics of the reciprocal relation between

167

team confidence and team performance.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

8

168

Although previous work on the relation between team confidence and team

169

performance revealed inconsistent results, most studies demonstrated a positive reciprocal

170

relation between both constructs; the more confident players were, the better they performed,

171

and vice versa (e.g., Myers, Paiement, et al., 2004; Stajkovic et al., 2009). Bandura (1997, p.

172

67) added that the relation between team confidence and performance is revealed most

173

accurately when both constructs are measured in close temporal proximity. Therefore, we

174

expected our results to demonstrate positive reciprocal relations between both types of team

175

confidence (i.e., (a) collective efficacy and (b) team outcome confidence) and team

176

performance. More specifically, we hypothesized that players‟ team confidence before the

177

game would be positively correlated with the perceived team performance in the first half

178

(H1a,b). Likewise, we hypothesized players‟ team confidence during half-time to be

179

positively correlated with the perceived team performance in the second half (H2a,b). On the

180

other hand, we also expected the perceived team performance during the first half to be a

181

significant predictor of players‟ team confidence during half-time (H3a,b). Finally, we

182

hypothesized the perceived team performance during the second half to be positively

183

correlated with players‟ team confidence after the game (H4).

184 185 186

Methods Recruitment In Study 1, the coaches of 13 Flemish soccer teams were invited via e-mail to

187

participate in our field study. Ten teams agreed to participate, leading to a response rate of

188

77%. In Study 2, a similar approach was maintained, resulting in a response rate of 67% and

189

again 10 participating teams. The most frequently cited reason for non-participation was the

190

refusal by the coach to allow measurements before the game or during half-time in order to

191

maintain the concentration of the players. There was no overlap in the samples of Study 1 and

192

Study 2.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 193

9

Before the warming-up, players and coaches were informed in detail about when the

194

different parts of the questionnaire had to be completed. The researcher was present in the

195

locker room to answer any questions. The APA ethical standards were followed in the

196

conduct of the study and players could withhold their participation at any time. No rewards

197

were given for participation in the study. Informed consent was obtained from all participants

198

and confidentiality was guaranteed.

199

Participants

200

Study 1. Ten soccer teams participated in the present study, including 134 male

201

players. Seven teams played at U17 regional level (i.e., youth teams playing at regional level

202

and only including players younger than 17 years old at the start of the season), two teams at

203

U17 provincial level, and one team at U19 national level. The players were on average 15.9

204

years old (SD = 0.8), had an average soccer experience of 9.5 years (SD = 2.4 years) of which

205

6.2 years in their current team (SD = 3.7 years). All participants filled out the questionnaires,

206

once before the game (i.e., before the warming-up) and both at the start and at the end of the

207

half-time break.

208

Study 2. This study also involved 10 teams, containing 125 male players. Seven teams

209

played at U17 regional level, one team at U21 regional level, and two teams participated in

210

the regional competition for adults. Participants were on average 17.3 years old (SD = 3.6),

211

played soccer for 10.0 years on average (SD = 4.7) of which 7.5 years in their current team

212

(SD = 4.5).

213

Measures

214

Team confidence. In line with previous research (Collins & Parker, 2010; Feltz &

215

Chase, 1998), Fransen, Kleinert, and colleagues (2014) conceptually distinguished between

216

outcome-oriented team confidence and process-oriented collective efficacy. We adopted this

217

conceptualization in our research, and assembled both concepts under the general term „team

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

10

218

confidence‟. Each study assessed both forms of team confidence at three different time points.

219

Study 1 assessed team confidence (i.e., both collective efficacy and team outcome confidence)

220

before the warming-up, at the beginning of half-time, and at the end of half-time. Study 2

221

assessed players‟ team confidence before the warming-up, at the beginning of half-time, and

222

after the game. Because there was no break between the warming-up and the start of the

223

game, the nearest moment at which players‟ team confidence could be measured was right

224

before the warming-up. As such, previous recommendations to measure team confidence at

225

least within 24h prior to the performance were taken into account (Feltz & Lirgg, 2001).

226

For the measurement after the game, each of the items began with the stem “If you

227

would compete once more against the same team, to what extent do you believe that your

228

team, during this new game, would …” The hypothetical situation of playing against the same

229

opponent was believed to be the most valid measure, because of its similarity with the

230

previous measures of team confidence before and during the game. If we had measured

231

players‟ team confidence after the game with regard to the next game (i.e., competing against

232

a different opponent), the ranking of that specific opponent could have led to a biased

233

response.

234

Collective efficacy. The Collective Efficacy Questionnaire for Sports (CEQS; Short,

235

Sullivan, & Feltz, 2005) included five subscales; Ability (e.g., “play more skillfully than the

236

opponent”), Effort (e.g., “demonstrate a strong work ethic”), Persistence (e.g., “persist when

237

obstacles are present”), Preparation (e.g., “devise a successful strategy”), and Unity (e.g.,

238

“keep a positive attitude”). Each of the items began with the stem “To what extent do you

239

believe that, during the upcoming game period, your team has the abilities to …” Fransen and

240

colleagues (2014) conducted an exploratory factor analysis which revealed that the CEQS

241

consisted of two factors; (1) the Ability subscale of the CEQS, and (2) the other four

242

subscales of the CEQS (i.e., Effort, Persistence, Preparation, and Unity). This factor analysis

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

11

243

demonstrated that the Ability subscale focused on the confidence in outplaying the opponent,

244

and as such is outcome-oriented, in contrast to the process-oriented nature of collective

245

efficacy, as originally defined by Bandura (1997). Therefore, in the present research, we will

246

focus on the subscales of Effort, Persistence, Preparation, and Unity that have been shown to

247

represent a valid measure of process-oriented collective efficacy (Fransen, Kleinert, et al.,

248

2014).

249

Both collective efficacy and team outcome confidence were measured at three

250

different time points in each study. Given the time constraints during half-time, it was not

251

possible to administer the full CEQS scale. As a consequence, to minimize the impact on the

252

team and to avoid concentration losses of the players, we only used the item with the highest

253

factor loading of each of the collective efficacy subscales (i.e., the example items as indicated

254

earlier). Participants assessed the items on a 7-point scale anchored by -3 (not at all confident)

255

and 3 (extremely confident). In the first study we administered the full CEQS scale before the

256

game as well. Our results revealed a strong correlation (r = .93; p < .01) between the 16-item

257

scale (including all items from subscales Effort, Persistence, Preparation, and Unity) and the

258

4-item scale (including only the highest loading item of each of these four subscales). The 4-

259

item scale revealed a high internal consistency throughout all measurement points (both in

260

Study 1 and Study 2, before, during, and after the game), demonstrated by Cronbach‟s alpha‟s

261

ranging from .81 to .91.

262

Team outcome confidence. In line with previous guidelines (Fransen, Kleinert, et al.,

263

2014), players assessed the item “To what extent do you believe that your team will win this

264

game?” on a 7-point scale anchored by -3 (not at all confident) and 3 (extremely confident).

265

Performance. Previous studies that examined the relation between team confidence

266

and performance mostly used objective measures such as scoring percentage, number of

267

turnovers, or game outcome to measure the team‟s performance (Feltz & Lirgg, 1998; Myers,

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

12

268

Paiement, et al., 2004; Watson, Chemers, & Preiser, 2001). However, Raglin and Morgan

269

(1988) pointed to the advantages of subjective measures of performance. These subjective

270

measures might be more accurate because they can account for performance indicators that

271

objective measures such as the game outcome cannot. To measure the team‟s performance,

272

we assessed players‟ subjective perceptions of the team‟s performance during half-time and

273

after the game. More specifically, players assessed the item “How well did your team play

274

during the previous half?” on a 7-point scale anchored by -3 (very bad) and 3 (very well). By

275

evaluating players‟ perceptions of the quality of their team‟s play, the present measure

276

focuses on the process, rather than on the outcome.

277

Data Analysis

278

The obtained data were analyzed with Stata version 13. For both Study 1 and Study 2,

279

the means, standard deviations, and bivariate correlations among collective efficacy, team

280

outcome confidence, and team performance measures were calculated. Due to the nesting of

281

the players within teams, we also calculated for each variable the proportion of variance

282

attributed to the team level.

283

Subsequently, the hypothesized relations were tested via structural equation modeling

284

using the maximum likelihood estimation method. The fit of the models was assessed using

285

the chi-square fit statistic (χ²), the goodness of fit index (GFI), the non-normed fit index

286

(NNFI), and the standardized root mean squared residual (SRMR). A non-significant χ²

287

indicates a good fit of the data to the proposed model. Incremental fit indices (GFI and NNFI)

288

had to be larger than 0.95. The SRMR, an absolute fit index had to be smaller than 0.06 to

289

accept a good fit (Hu & Bentler, 1999).

290

In addition, the hypothesized structural equation models were analyzed in a multilevel

291

analysis to test the variance in intercepts and slopes that might be attributed to the nesting of

292

players within teams. This was done by comparing the likelihood ratios of the fixed model

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

13

293

with a χ² estimation when allowing for random intercepts, and a χ² estimation when allowing

294

for random slopes.

295 296

Results Descriptive statistics and correlations among the variables are provided in Table 1 for

297

both studies. The measurements of players‟ team confidence before the game, during the

298

game, and after the game were only moderately correlated, illustrating the dynamic nature of

299

team confidence and its variation within a single game. This was found for collective efficacy

300

(r = .42 in Study 1; r = .27 – .67 in Study 2) as well as for team outcome confidence (r = .48

301

in Study 1; r = .36 – .48 in Study 2). Furthermore, the correlations between process-oriented

302

collective efficacy and outcome-oriented team outcome confidence before the game (.46 in

303

Study 1; .49 in Study 2) are clearly lower than the correlations between both constructs during

304

and after the game (respectively .75 and .82 in Study 1; .67 and .69 in Study 2). In addition, it

305

is noteworthy that these correlations were only moderately correlated at all three measurement

306

time-points (i.e., before, during, and after the game), indicating that collective efficacy and

307

team outcome confidence, although related, are two distinct constructs.

308

When the total variance was partitioned into variance at the team level and into

309

variance at the individual level, the results revealed that the proportion of variance at the team

310

level ranged between 20% and 57% in Study 1 and between 8% and 62% in Study 2. For

311

every variable the likelihood ratios with and without the team-level variance component was

312

significantly different (p < .05). This finding indicates that for all variables the variance

313

proportion at the team level cannot be disregarded. The team variance proportions are

314

provided in the first column of Table 2.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

14

315

Table 1. Means, standard deviations, and correlations across all measures of team outcome

316

confidence (TOC), collective efficacy (CE), and players’ perceived team performance for both

317

studies.

Variable Study 1 1. TOC before the game 2. TOC start half-time 3. TOC end half-time 4. CE before the game 5. CE start half-time 6. CE end half-time 7. Team performance first half 8. Team performance second half Study 2 1. TOC before the game 2. TOC half-time 3. TOC after the game 4. CE before the game 5. CE half-time 6. CE after the game 7. Team performance first half 8. Team performance second half * 318 p < .05; **p < .01

M

SD

2.28 1.98 2.02 1.87 2.09 2.12 .74 1.22 1.72 1.75 1.81 1.62 1.84 1.79 .45 .86

1

2

3

4

5

6

7

8

1.11 1 .48** .52** .46** 1.18 1 .81** .37** 1.15 1 .31** .94 1 .93 .89 1.27 1.36

.41** .75** .72** .42** 1

.47** .73** .82** .40** .81** 1

-.10 .28** .23** .06 .33** .27** 1

.16 .39** .39** .18* .41** .44** .40** 1

1.26 1 .36** .37** .49** 1.10 1 .48** .26** 1.20 1 .36** .94 1 .91 .97 1.67 1.53

.32** .67** .49** .34** 1

.28** .53** .69** .27** .67** 1

.01 -.13 .38** .01 .20** .13 .15 -.03 .31** .25** .29** .34** 1 .18 1

319 320

Table 2. Variance partition coefficients of team outcome confidence (TOC), collective

321

efficacy (CE), and players’ perceived team performance for both studies. Null model

Structural equation model

Variance at team level

Explained Explained variance Unexplained variance at team at individual level (residual) level (%) (%) variance (%)

Study 1 TOC before the game

57% *

-

-

-

26%

*

3%

34%

63%

TOC end half-time

26%

*

0%

69%

31%

CE before the game

34% *

-

-

-

CE start half-time

23% *

8%*

25%

67%

CE end half-time

20% *

0%

66%

34%

Performance 1st half

38% *

-

-

-

Performance 2nd half (a)

39% *

23%*

28%

49%

Performance 2nd half (b)

39% *

25%*

26%

49%

TOC start half-time

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

15

Study 2 TOC before the game

-

-

-

*

1%

26%

73%

11% *

0%

32%

68%

CE before the game

8% *

-

-

-

CE half-time

9% *

7%

17%

76%

CE end of the game

18% *

0%

48%

52%

Performance 1st half

62% *

-

-

-

Performance 2nd half (a)

59% *

61%*

7%

32%

Performance 2nd half (b)

59% *

62%*

4%

34%

TOC half-time TOC end of the game

322

28% *

*

9%

Team-level variance component adds significantly to the model‟s likelihood ratio (p < .05).

323 324

Study 1

325

For Study 1, the hypothesized relations between both types of team confidence (i.e.,

326

collective efficacy and team outcome confidence) and the team‟s perceived performance in

327

the first and second half were modeled in a structural equation model, which is shown in

328

Figure 1 for collective efficacy and Figure 2 for team outcome confidence. The dotted

329

pathways were hypothesized, but failed to show significant regression weights at the p < .05

330

level. Additionally, modification indices suggested that subsequent assessments of collective

331

efficacy, team outcome confidence, and team performance were also directly predicted by

332

their prior measures. These additional suggested pathways were added and both models

333

provided evidence of a good fit to our data.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

(.23).28 Team performance 5

.31

Team performance first half

second half .32

n.s.

.36 (.08).25

334

Collective efficacy before the game

.39

16

Collective efficacy start half-time

(.00).66 .81

Collective efficacy end half-time

335

Figure 1. The structural model of Study 1 for the reciprocal relation between players’

336

process-oriented collective efficacy and their perceived team performance. All regression

337

coefficients are standardized, significant (p < .001), and presented along the pathways. The

338

proportion of predicted variance is noted above the predicted variables. The team-level

339

variance is shown between parentheses. Goodness-of-fit indices are: χ²(df = 4) = 3.73, p =

340

.44, CFI = 1.00, NNFI = 1.00, and SRMR = .03.

341 (.25) .26 .35

Team performance first half

Team performance second half

.33 (.03) .34

n.s. .50

342

Team outcome confidence before the game

Team outcome confidence start half-time .18

.31 .73

(.00) .69

Team outcome confidence end half-time

343

Figure 2. The structural model of Study 1 for the reciprocal relation between the players’

344

outcome-oriented team outcome confidence and their perceived team performance. All

345

regression coefficients are standardized, significant (p < .01), and presented along the

346

pathways. The proportion of predicted variance is noted above the predicted variables. The

347

team-level variance is shown between parentheses. Goodness-of-fit indices are: χ²(df = 3) =

348

1.51, p = .68, CFI = 1.00, NNFI = 1.02, and SRMR = .02.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 349

17

Partial support for the reciprocal relations between players‟ team confidence and

350

perceptions of the team‟s performance was found. In contrast to H1, no significant relation

351

was found between the team‟s confidence before the game and its performance during the

352

first half (according to the perceptions of the players), neither for collective efficacy (H1a; p =

353

.99), nor for team outcome confidence (H1b; p = .46). By contrast, the measures obtained

354

during games confirmed the reciprocal relation between players‟ team confidence and the

355

team‟s performance; a positive relation was found between the team‟s confidence at the end

356

of half-time and the team‟s perceived performance in the second half (for collective efficacy

357

(H2a): β = .36, p < .001; for team outcome confidence (H2b): β = .31, p < .001). These

358

findings confirm H2; the more confident the players were in the capacities of their team

359

during half-time, the better they perceived their performance in the second half. Furthermore,

360

in line with H3, a positive relation appeared between the team‟s perceived performance

361

during the first half and both types of players‟ confidence at the beginning of half-time (for

362

collective efficacy (H3a): β = .32, p < .001; for team outcome confidence (H3b): β = .33, p <

363

.001). The better the team performed, the more confident the players were (a) in the capacities

364

of their team to successfully complete the process-oriented tasks and (b) in winning the game.

365

Study 2

366

Similar to the analysis in Study 1, the reciprocal relations between players‟ team

367

confidence and perceived team performance were tested in a structural equation model but

368

Study 2 included a measurement of team confidence after the game. Again, dotted lines

369

indicate that the predicted relations were not significant (p > .05). As suggested by

370

modification indices, subsequent measures of the same construct were connected. The

371

resulting models, including the standardized regression path coefficients and the proportions

372

explained variance, are shown in Figure 3 for collective efficacy and Figure 4 for team

373

outcome confidence. Both models showed a good fit to our data.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

18

(.61) .07 n.s.

Team performance first half n.s.

.28

Team performance second half .20

.19 (.00) .48

(.07) .17

374

Collective efficacy before the game

.29

Collective efficacy during half-time

.63

Collective efficacy after the game

375

Figure 3. The structural model of Study 2 for the reciprocal relation between the players’

376

process-oriented collective efficacy and their perceived team performance. All regression

377

coefficients are standardized, significant (p < .01), and presented along the pathways. The

378

proportion of predicted variance is noted above the predicted variables. The team-level

379

variance is shown between parentheses. Goodness-of-fit indices are: χ²(df = 3) = 4.40, p =

380

.22, CFI = .99, NNFI = .95, and SRMR = .04.

381 (.62) .04 .22

Team performance first half .37

n.s. (.01) .26

n.s. .34

382

Team outcome confidence before the game

Team performance second half

Team outcome confidence half-time .26

.16 .39 (.00) .32 Team outcome confidence after the game

383

Figure 4. The structural model of Study 2 for the reciprocal relation between the players’

384

outcome-oriented team outcome confidence and their perceived team performance. All

385

regression coefficients are standardized, significant (p < .05), and presented along the

386

pathways. The proportion of predicted variance is noted above the predicted variables. The

387

team-level variance is shown between parentheses. Goodness-of-fit indices are: χ²(df = 2) =

388

1.12, p = .57, CFI = 1.00, NNFI = 1.06, and SRMR = .02.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 389

19

In contrast to H1, but in line with the findings of Study 1, no significant regression

390

was found between both forms of players‟ team confidence before the game and the team‟s

391

perceived performance during the first half (for collective efficacy p = .22; for team outcome

392

confidence p = .84). Our expectation that the team‟s confidence during half-time would be a

393

predictor of the team‟s perceived performance during the second half (H2) was confirmed for

394

collective efficacy (H2a; β = .20, p < .01), but not for team outcome confidence (H2b; p =

395

.40). In other words, players‟ confidence in the game‟s outcome did not affect the team‟s

396

performance in the next half. However, players who were confident during half-time in the

397

team‟s abilities to demonstrate a strong work ethic, to persist when encountering difficulties,

398

to devise a successful strategy, and to keep a positive attitude, perceived their team as

399

performing better in the second half.

400

In line with H3 and the findings of Study 1, a positive relation existed between the

401

team‟s perceived performance during the first half and players‟ team confidence during half-

402

time (for collective efficacy (H3a) β = .28, p < .01; for team outcome confidence (H3b) β =

403

.37, p < .05). Specifically in Study 2, H4 was confirmed by demonstrating a significant

404

positive association between the team‟s perceived performance during the second half and the

405

players‟ team confidence after the game (for collective efficacy (H4a) β = .19, p < .01; for

406

team outcome confidence (H4b) β = .16, p < .05). In other words, perceptions of a better team

407

performance during the previous half went hand in hand with a stronger confidence in the

408

team‟s abilities to fulfill the required processes and to win the game.

409

Multilevel Analysis

410

Testing the same models in a generalized structural model with random intercepts

411

across teams revealed a significant proportion of variance at team level (for collective efficacy

412

in Study 1:χ² (df = 2) = 22.99, p < .001; for collective efficacy in Study 2: :χ² (df = 2) =

413

89.79, p <.001; for team outcome confidence in Study 1: χ² (df = 2) = 22.13, p < .001; and

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

20

414

for team outcome confidence in Study 2:χ² (df = 2) = 77.66, p < .001). However, an

415

intercept by intercept analysis revealed that the initial values of collective efficacy and team

416

outcome confidence predicted more variance of respective subsequent measures than the

417

portion of variance at team level. For these measures, the variance at team level decreased as

418

prior measures were taken into account. Only for the team‟s performance in the second half,

419

in both models in both studies, a substantial random team effect remained. The predicted

420

variances at team and individual level are provided in Table 2.

421

Adding random slope effects to the random intercept models failed to show significant

422

added variance (all p > .05). An exception was found with respect to the pathway from

423

collective efficacy before the game to collective efficacy during half-time in Study 2 (χ² (df

424

= 2) = 9.05, p < .05). This random slope effect of .08 did not covary significantly with the

425

respective random intercept coefficient (p > .05) and was the only significant random slope

426

detected among all regressions in the four models.

427

Discussion

428

The present research extended previous research in two ways. First, within a field

429

context, players‟ team confidence was assessed in a quantitative way, not only before and

430

after the game, but for the first time also during the game. Our findings highlight the dynamic

431

nature of team confidence, demonstrated by the variation of players‟ team confidence within a

432

single game. This observation contrasts with previous assumptions that team confidence prior

433

to the competition is relatively stable throughout the competition (Myers et al., 2007). Second,

434

we conceptually distinguished between process-oriented collective efficacy and outcome-

435

oriented team outcome confidence and examined their relation with perceived team

436

performance. Our findings provide partial support for the reciprocal relation between players‟

437

team confidence (including both team outcome confidence and collective efficacy) and

438

players‟ perceptions of the team‟s performance.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 439

21

Neither within Study 1, nor within Study 2, a significant relation emerged between

440

players‟ team confidence before the game (both collective efficacy and team outcome

441

confidence) and the team‟s perceived performance during the first half (H1). With regard to

442

the second half of the game (H2), inconsistent results were found for team outcome

443

confidence; Study 1 revealed that players‟ team outcome confidence during half-time

444

positively predicted the perceptions of the team‟s performance during the second half, but this

445

was not confirmed by Study 2. Regarding collective efficacy, both studies provided support

446

for a significant association between players‟ collective efficacy during half-time and the

447

team‟s perceived performance during second half. The abovementioned results thus partially

448

confirmed Hypotheses 1 and 2 stating that players‟ team confidence is a significant predictor

449

of the team‟s performance in the subsequent half.

450

Having confidence in the team‟s abilities to successfully perform the required process

451

(i.e., collective efficacy) was more strongly associated with the team‟s subsequent

452

performance perceptions than the confidence in winning the game (i.e., team outcome

453

confidence). A plausible underpinning of this finding is the concordance between the

454

measures of team confidence and the way in which performance was measured. As outlined

455

by Myers, et al. (2007), assessments of team confidence and team performance are concordant

456

when both tap similar capabilities (e.g., confidence in winning the game and performance

457

measured by game outcome). The relation between confidence and performance is expected

458

to be the strongest when the two constructs are not only measured in close temporal

459

proximity, but when they are also concordant (Bandura, 1997).

460

In our study, the performance was measured by players‟ subjective perceptions of the

461

overall team performance. By evaluating players‟ perceptions of the quality of their team‟s

462

play, the present measure focuses on the process, rather than on the outcome. Therefore, it can

463

be derived that the measure of collective efficacy (representing the confidence in the

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

22

464

processes underlying the performance) is more concordant with the performance measure that

465

we used than is the confidence in winning the game. For example, if a team plays against a

466

weakly performing opponent, it is likely that players will not base their performance ratings

467

predominantly on the game outcome, but instead use a process-based evaluation to rate

468

whether their team has played well.

469

The different findings for the first and second half reflect the inconsistency found in

470

previous literature. Although some studies demonstrated that team confidence judgments

471

taken prior to the competition are predictive of team performance throughout the competition

472

(Chou, Yu, & Chi, 2010; Edmonds et al., 2009; Feltz & Lirgg, 1998; Myers, Paiement, et al.,

473

2004; Myers et al., 2007), other studies did not find such a link (MacLean & Sullivan, 2003;

474

Watson et al., 2001). Chen et al. (2002) conducted both a laboratory study and a field study to

475

test this relation. Although the laboratory study revealed that collective efficacy positively

476

predicted team performance, this relation was not replicated in the field sample. These

477

findings are consistent with previous meta-analytic studies on self-efficacy (Stajkovic &

478

Luthans, 1998), which suggest that efficacy beliefs predict performance more strongly in

479

laboratory settings than in field settings. A plausible rationale for this finding might reside in

480

the situational unpredictability of the surrounding circumstances in field studies, compared to

481

the highly controlled circumstances in laboratory experiments. As Bandura (1997, p. 64)

482

stated “if one does not know what demands must be fulfilled in a given endeavor, one cannot

483

accurately judge whether one has the requisite abilities to perform the task.” The fact that the

484

present research includes two field studies may explain why no significant effect was found

485

between players‟ team confidence before the game and the perceived performance during the

486

first half.

487 488

However, it should be considered that players‟ team confidence before the game is based on general impressions (such as the team‟s playing level in previous games, the ranking

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

23

489

of the opponent, etc.), whereas players‟ team confidence during half-time is the result of much

490

more concrete experiences during the game (e.g., present-day playing level of the own team

491

and of the opponent). This difference might explain why the team confidenceperformance

492

relation was not found for the first half, but did emerge in the second half.

493

Another plausible reason for this discrepancy in the relation between team confidence

494

and performance relates to the time between the measurements. Previous research (Bandura,

495

1997; Myers & Feltz, 2007) stated that the relation between team confidence and performance

496

is revealed most accurately when both constructs are measured in close temporal proximity.

497

The time lapse between the measurement of team confidence before the game (i.e., before the

498

warming-up) and the team‟s perceived performance in the first half allowed for intervening

499

experiences that may have impacted on the team‟s confidence, such as the pre-game speech of

500

the coach, the team appearance of the opponent during the warming-up, or the cheering of the

501

audience (Ronglan, 2007; Vargas-Tonsing & Bartholomew, 2006). The much smaller time

502

lapse between half-time and the team‟s performance during second half may have accounted

503

for a more accurate measure of players‟ team confidence during half-time, resulting in a

504

significant team confidence–performance relation within the game.

505

The second aim of our research was to examine whether previous perceptions of the

506

team‟s performance were a significant predictor of players‟ team confidence. The present

507

findings provided empirical support for that hypothesis. More specifically, Study 1 and Study

508

2 demonstrated a significant relation between the perceived team performance during the first

509

half and both types of players‟ team confidence during half-time (H3). Furthermore, Study 2

510

added evidence for a significant relation between the perceived team performance during

511

second half and both forms of players‟ team confidence after the game (H4). These results are

512

consistent with Bandura‟s theory (1997) that points to prior performance as one of the most

513

important sources of team confidence. Several studies confirmed this statement and revealed

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

24

514

that as teams performed better, the more confident they became concerning the abilities of

515

their team (Feltz & Lirgg, 1998; Heuze, Raimbault, & Fontayne, 2006; Myers, Paiement, et

516

al., 2004; Stajkovic et al., 2009; Stanimirovic & Hanrahan, 2004).

517

Although Myers and Feltz (2007) recommended multilevel modeling as the optimal

518

framework for analyzing collective efficacy data, their meta-analysis demonstrated that

519

previous studies rarely used a multilevel approach. Submitting meaningfully nested observed

520

data to multilevel modeling is seen as the most efficient, most unbiased, and most appropriate

521

way to analyze this type of data (Raudenbush & Bryk, 2002). In contrast to these

522

recommendations, most researchers have focused on either the individuals within groups or

523

the group as a whole, but seldom on both (Moritz & Watson, 1998).

524

In the present manuscript, the data of both studies were analyzed by a multilevel

525

approach. Our findings revealed that the variance of the measured constructs was explained

526

both at the individual level (i.e., within-team level) and at the team level (i.e., between-team

527

level). The regression weights between the different constructs did not vary at team level,

528

indicating that the impact of team confidence on perceived performance and vice versa is

529

similar for every individual player regardless of the team.

530

The variance of players‟ perceptions of their team‟s performance was mainly

531

explained at team level, both for first and second half. With regard to collective efficacy and

532

team outcome confidence, the variance explained at team level decreased with time; although

533

a significant part of the variance of both constructs before the game was explained at team

534

level, during the game the individual perception was the factor that explained most variance.

535

This finding implies that no team effects emerged during the game (e.g., no impact of a

536

motivational speech of the coach directed at the whole team).

537

Because collective efficacy was originally considered as a group level construct, many

538

studies have used an approach that assesses each player‟s belief in the team‟s capabilities as a

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

25

539

whole and then aggregates these individual measures to the team level (Myers, Feltz, & Short,

540

2004; Myers, Paiement, et al., 2004). Although Bandura (2000) assumed that this aggregated

541

collective efficacy estimate is a better predictor of team performance within highly interactive

542

tasks, the present research suggests that, during the game, the focus should be on the

543

individual perceptions of team confidence, rather than on the aggregated team perception.

544

When interpreting the present findings, it is worth considering the strengths and

545

weaknesses of our study approach. A major strength of this research is that for the first time

546

players‟ team confidence was assessed not only before and after the game, but also during the

547

game. This in-game measurement allowed us to capture the dynamic nature of players‟ team

548

confidence within the game. Although Myers et al. (2007) assumed that players‟ team

549

confidence prior to the competition may be relatively stable during the performance, the

550

moderate correlations between team confidence before, during, and after the game obtained in

551

the present studies reveal that team confidence did fluctuate during the game. This finding

552

emphasizes the need to examine team confidence as a dynamic construct instead of as a trait-

553

like characteristic with a strong cross-temporal stability.

554

A second strength of the present study is that we conceptually distinguished between

555

two forms of team confidence in our two studies; process-oriented collective efficacy and

556

outcome-oriented team outcome confidence. Although most relations were consistent across

557

both forms, an important difference was demonstrated in Study 2; in contrast to team outcome

558

confidence, collective efficacy during half-time was shown to be a significant predictor for

559

the team‟s performance in the second half. The team‟s belief in the process (i.e., collective

560

efficacy) is much more controllable than the team‟s belief to win (i.e., team outcome

561

confidence), which is more susceptible to external factors such as the opponent, dubious

562

referee decisions, or a lucky goal. Given its stronger link with the subsequent team

563

performance, coaches and athlete leaders should primarily focus on enhancing players‟

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

26

564

collective efficacy, which in turn may foster the team‟s outcome confidence (Fransen, Coffee,

565

Vanbeselaere, Slater, De Cuyper, & Boen, 2014).

566

In addressing the limitations of the present research, several opportunities for future

567

research emerge. First, although the team‟s performance was demonstrated to be a significant

568

predictor of players‟ team confidence, it should be noted that the production of team

569

confidence is an interpersonal process, brought about not only by perceptions of previous

570

performances, but also by persuasive actions of the coach or athlete leaders, by motivational

571

and tactical communication within the team, and by the enthusiasm expressed by the team

572

members (Fransen, Coffee, et al., 2014; Fransen et al., 2012; Ronglan, 2007). Future research

573

may investigate how these behaviors affect players‟ team confidence within a game and as

574

such the subsequent team performance.

575

Second, we chose to assess players‟ subjective perception of the team‟s performance.

576

Although Raglin and Moran (1988) pointed to the advantages of these subjective measures of

577

performance (e.g., more accurate because they can account for performance indicators that

578

objective measures, such as game outcome, cannot), some limitations should be denoted. Self-

579

serving bias for example can distort these performance perceptions by the need to maintain

580

and enhance self-esteem. In this regard, players are more likely to attribute a winning game to

581

their own abilities (i.e., internal attribution), while blaming a defeat to the circumstances (i.e.,

582

external attribution). This self-serving bias would involve that the subjective perceptions of

583

performance represent an overestimation of the actual performance.

584

Although our subjective measures of performance varied between .45 and 1.22 on a

585

scale from -3 to 3, and as such did not reflect a ceiling effect, examining the in-game relation

586

between team confidence and both subjective and objective measures of performance might

587

be a fruitful line for further research. In this regard, objective performance measures should

588

not only focus on the outcome, but should also include process indicators. Future research

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

27

589

could use the recently developed technological devices and mathematical methods to analyze

590

the performance of soccer players (Clemente, Couceiro, Martins, Mendes, & Figueiredo,

591

2013; Couceiro, Clemente, Martins, & Tenreiro Machado, 2014). Such performance measures

592

can capture both technical and tactical performance, indicated by factors such as ball

593

possession, the covered distance, etc.

594

Third, constrained by practical feasibility, we included only one measurement point

595

within the game, namely during half-time. Future research may explore the dynamic relation

596

between team confidence and performance even further by including more measurement

597

points within the game. Other team sports that are characterized by multiple breaks within a

598

game, such as volleyball or basketball, might be more appropriate to reach this aim. When

599

aiming for even more dynamic in-game measurements, using continuous observations instead

600

of questionnaires to measure team confidence would be an important step forward to capture

601

the dynamic in-game relation between team confidence and performance (Fransen, Kleinert,

602

et al., 2014).

603

Fourth, given the time constraints during half-time, it was not possible to administer

604

the full CEQS scale. Instead, we used the short version of the CEQS, which has lower

605

psychometric qualities. However, it should be noted that this questionnaire assesses five

606

specific behaviors that might not capture the key processes underlying the team performance.

607

Therefore, future research should establish whether the same results are observed when using

608

a collective measure that includes the most important game competencies specific for a given

609

sport (e.g., the measures used in Myers, Feltz, et al., 2004; Myers, Paiement, et al., 2004).

610

Fifth, with regard to the participants in our study, we mainly assessed older youth

611

players. Future research should examine whether our findings can be generalized to other age

612

groups and other competition levels. With regard to age, it is likely that the team confidence

613

of mature players is more stable over time. Furthermore, in high-level teams, the team

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

28

614

confidence of the different players within a team could be more homogeneous. A plausible

615

underlying reason for this homogeneity is that in high-level teams the coach is expected to

616

have a higher impact on the players, thereby influencing the team confidence on the team

617

level. Furthermore, high-level players spend more training time together in which the

618

underlying processes for performance are practiced. As such, it is likely that high-level teams

619

share a common confidence in their abilities to perform these processes successfully. As a

620

consequence, we expect that more variance of collective efficacy and team outcome

621

confidence is explained at team level in high-level teams than in low-level teams.

622

In addition, only soccer players participated in our study. Considering that the

623

outcome in soccer is more unpredictable and susceptible to external factors, such as a lucky

624

goal or a dubious referee decision, it remains to be determined whether our findings apply to

625

other sports as well. For instance, in games such as volleyball and basketball, in which the

626

scoring range is much higher, and as such, the game outcome is more controllable and

627

represents the playing level of both teams better, future research should examine whether

628

team confidence relates similarly to performance in these sports as was the case in soccer.

629

Another fruitful line for future research pertains to the stability of players‟ team

630

confidence. Although many studies have assessed players‟ team confidence, the strength of

631

this confidence, or in other words, the stability of this confidence over time, has only rarely

632

been measured. However, considerable individual differences might exist regarding the

633

stability of one‟s team confidence; some players‟ team confidence is strong, in the sense that

634

this confidence is able to resist even the strongest pressures to change (such as being behind

635

in the game, a teammate‟s injury, etc.). On the other hand, if a player‟s team confidence is

636

unstable and vulnerable to situational pressures, overconfidence at the start of the game might

637

lead to a collapse (both in confidence and performance) if the team is performing worse than

638

expected. Therefore, in line with literature on attitudes (Krosnick & Abelson, 1992), further

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

29

639

research could include a measure for the strength or stability of team confidence over time,

640

and investigate the link with performance.

641

There are a number of practical implications that could be considered by coaches,

642

sport psychologists, and sports teams. First, the only moderate correlations of collective

643

efficacy before, during, and after the game demonstrate that collective efficacy is amenable to

644

change. In this regard, it is important to note that the multilevel analyses of the present study

645

showed that the variance of team confidence during the game is mainly explained at the

646

individual level. Therefore, coaches should strive to enhance each player‟s team confidence in

647

an individualized way. Based on the present findings, such an individual approach is likely to

648

be more effective than a motivational speech for the whole group.

649

Second, our findings did not demonstrate a significant relation between players‟ team

650

confidence before the game and their playing level during first half. In line with the

651

abovementioned comments on team confidence stability, it might be better for coaches to

652

strive for a realistic, but stable team confidence before the game, for instance by strengthening

653

players‟ confidence in their team‟s tactical game plan. As such, unrealistic overconfidence at

654

the start of the game can be avoided, thereby reducing the chances on confidence collapses

655

during the game if the team‟s performance falls short. Because our findings suggest that a

656

players‟ team confidence during half-time is a positive predictor of the team‟s performance in

657

the second half, it seems important for coaches to create a team confidence that is not only

658

high, but also stable throughout the game.

659

Not only coaches, but also athlete leaders within the team play a key role in enhancing

660

the team‟s confidence and preventing downward efficacyperformance spirals (Lindsley,

661

Brass, & Thomas, 1995). Several studies pointed out that leaders who display confidence are

662

more likely to enhance collective efficacy among their teammates (Fransen et al., 2012;

663

Moritz & Watson, 1998; Vargas-Tonsing et al., 2004; Zaccaro, Rittman, & Marks, 2001).

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

30

664

Furthermore, verbal persuasion can be used as an effective form to increase players‟ team

665

confidence (Vargas-Tonsing et al., 2004). Ronglan (2007) added that team confidence

666

building might be facilitated if key players use their leader status to affect their teammates‟

667

confidence positively. As such, an important task for coaches is to make their athlete leaders

668

aware of their potential and responsibility as role models in the team.

669

In conclusion, the current manuscript provided a deeper insight into the dynamics of

670

the reciprocal relation between team confidence and perceived performance within soccer

671

games. Given the fact that both process-oriented collective efficacy and team outcome

672

confidence are dynamic processes that can be controlled by coach and players, the present

673

findings open new avenues to optimize the team‟s performance.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 674

References

675

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman.

676

Bandura, A. (2000). Exercise of human agency through collective efficacy. Current

677 678

31

Directions in Psychological Science, 9, 75-78. Chen, G., Webber, S. S., Bliese, P. D., Mathieu, J. E., Payne, S. C., Born, D. H., & Zaccaro,

679

S. J. (2002). Simultaneous examination of the antecedents and consequences of

680

efficacy beliefs at multiple levels of analysis. Human Performance, 15(4), 381-409.

681

Chou, Y. H., Yu, C. A., & Chi, L. K. (2010). The relationships between coaching efficacy,

682

collective efficacy, team performance and satisfaction in youth softball teams. Journal

683

of Sport & Exercise Psychology, 32, S151-S151.

684

Clemente, F. M., Couceiro, M. S., Martins, F. M. L., Mendes, R., & Figueiredo, A. J. (2013).

685

Measuring tactical behaviour using technological metrics: Case study of a football

686

game. International Journal of Sports Science & Coaching, 8(4), 723-740. doi:

687

10.1260/1747-9541.8.4.723

688

Collins, C. G., & Parker, S. K. (2010). Team capability beliefs over time: Distinguishing

689

between team potency, team outcome efficacy, and team process efficacy. Journal of

690

Occupational and Organizational Psychology, 83(4), 1003-1023. doi:

691

10.1348/096317909x484271

692

Couceiro, M. S., Clemente, F. M., Martins, F. M. L., & Tenreiro Machado, J. A. (2014).

693

Dynamical stability and predictability of football players: The study of one match.

694

Entropy, 16, 645-674.

695

Edmonds, W. A., Tenenbaum, G., Kamata, A., & Johnson, M. B. (2009). The role of

696

collective efficacy in adventure racing teams. Small Group Research, 40(2), 163-180.

697

doi: 10.1177/1046496408328489

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

32

698

Feltz, D. L., & Chase, M. A. (1998). The measurement of self-efficacy and confidence in

699

sport. In J. L. Duda (Ed.), Advancements in sport and exercise psychology

700

measurement (pp. 63-78). Morgantown, WV: Fitness Information Technology.

701

Feltz, D. L., & Lirgg, C. D. (1998). Perceived team and player efficacy in hockey. Journal of

702 703

Applied Psychology, 83(4), 557-564. Feltz, D. L., & Lirgg, C. D. (2001). Self-efficacy beliefs of athletes, teams, and coaches. In R.

704

N. Singer, H. A. Hausenblas & C. M. Janelle (Eds.), Handbook of sport psychology

705

(2nd ed., pp. 340-361). New York: Wiley.

706

Fransen, K., Coffee, P., Vanbeselaere, N., Slater, M., De Cuyper, B., & Boen, F. (2014). The

707

impact of athlete leaders on team members‟ team outcome confidence: A test of

708

mediation by team identification and collective efficacy. The Sport Psychologist,

709

Manuscript in press. doi: http://dx.doi.org/10.1123/tsp.2013-0141

710

Fransen, K., Kleinert, J., Dithurbide, L., Vanbeselaere, N., & Boen, F. (2014). Collective

711

efficacy or team outcome confidence? Development and validation of the

712

Observational Collective Efficacy Scale for Sports (OCESS). International Journal of

713

Sport Psychology, 45(2), 121-137. doi: 10.7352/IJSP 2014.45.121.

714

Fransen, K., Vanbeselaere, N., Exadaktylos, V., Vande Broek, G., De Cuyper, B., Berckmans,

715

D., . . . Boen, F. (2012). "Yes, we can!": Perceptions of collective efficacy sources in

716

volleyball. Journal of Sports Sciences, 30(7), 641-649. doi:

717

10.1080/02640414.2011.653579

718

Greenlees, I. A., Graydon, J. K., & Maynard, I. W. (1999). The impact of collective efficacy

719

beliefs on effort and persistence in a group task. Journal of Sports Sciences, 17(2),

720

151-158.

721 722

Heuzé, J. P., Raimbault, N., & Fontayne, P. (2006). Relationships between cohesion, collective efficacy and performance in professional basketball teams: An examination

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 723

of mediating effects. Journal of Sports Sciences, 24(1), 59-68. doi:

724

10.1080/02640410500127736

725 726 727

33

Hodges, L., & Carron, A. V. (1992). Collective efficacy and group performance. International Journal of Sport Psychology, 23(1), 48-59. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure

728

analysis: Conventional criteria versus new alternatives. Structural Equation Modeling-

729

a Multidisciplinary Journal, 6(1), 1-55. doi: 10.1080/10705519909540118

730

Keshtan, M. H., Ramzaninezhad, R., Kordshooli, S. S., & Panahi, P. M. (2010). The

731

relationship between collective efficacy and coaching behaviors in professional

732

volleyball league of Iran clubs. World Journal of Sport Sciences, 3(1), 1-6.

733

Krosnick, J. A., & Abelson, R. P. (1992). The case for measuring attitude strength in surveys.

734

In J. M. Tanur (Ed.), Questions about questions: Inquiries into the cognitive bases of

735

surveys (pp. 177-203). New York: Russell Sage Foundation.

736

Lindsley, D. H., Brass, D. J., & Thomas, J. B. (1995). Efficacy-performance spirals: A

737

multilevel perspective. Academy of Management Review, 20(3), 645-678.

738

MacLean, D., & Sullivan, P. (2003). A season long case study investigation of collective

739

efficacy in male intercollegiate basketball. Athletic Insight: The Online Journal of

740

Sport Psychology, 5(3), 1-9.

741

Morgan, P. B. C., Fletcher, D., & Sarkar, M. (2013). Defining and characterizing team

742

resilience in elite sport. Psychology of Sport and Exercise, 14(4), 549-559. doi:

743

10.1016/j.psychsport.2013.01.004

744

Moritz, S. E., & Watson, C. B. (1998). Levels of analysis issues in group psychology: Using

745

efficacy as an example of a multilevel model. Group Dynamics: Theory, Research,

746

and Practice, 2(4), 285-298.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 747

34

Myers, N. D., & Feltz, D. L. (2007). From self-efficacy to collective efficacy in sport:

748

Transitional methodological issues. In G. Tenenbaum & R. C. Eklund (Eds.),

749

Handbook of sport psychology (3rd ed., pp. 799-819). Hoboken, NJ, US: John Wiley

750

& Sons Inc.

751

Myers, N. D., Feltz, D. L., & Short, S. E. (2004). Collective efficacy and team performance:

752

A longitudinal study of collegiate football teams. Group Dynamics: Theory, Research,

753

and Practice, 8(2), 126-138. doi: 10.1037/1089-2699.8.2.126

754

Myers, N. D., Paiement, C. A., & Feltz, D. L. (2004). Reciprocal relationships between

755

collective efficacy and team performance in women's ice hockey. Group Dynamics:

756

Theory, Research, and Practice, 8(3), 182-195. doi: 10.1037/1089-2699.8.3.182

757

Myers, N. D., Paiement, C. A., & Feltz, D. L. (2007). Regression team performance on

758

collective efficacy: Considerations of temporal proximity and concordance.

759

Measurement in Physical Education and Exercise Science, 11(1), 1-24.

760

Quinn, A. (2012). How to be an extraordinary athlete: The secrets to sporting success.

761 762 763 764 765 766 767 768

Melbourne, Australia: Quintessential Publishing. Raglin, J. S., & Morgan, W. P. (1988). Predicted and actual pre-competition anxiety in college swimmers. Journal of Swimming Research, 2, 5-7. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Newbury Park, CA: Sage. Ronglan, L. T. (2007). Building and communicating collective efficacy: A season-long indepth study of an elite sport team. The Sport Psychologist, 21(1), 78-93. Shearer, D. A., Holmes, P., & Mellalieu, S. D. (2009). Collective efficacy in sport: The future

769

from a social neuroscience perspective. International Review of Sport and Exercise

770

Psychology, 2(1), 38-53. doi: 10.1080/17509840802695816

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE

35

771

Short, S. E., Sullivan, P., & Feltz, D. (2005). Development and preliminary validation of the

772

collective efficacy questionnaire for sports. Measurement in Physical Education and

773

Exercise Science, 9(3), 181-202.

774 775 776 777 778

Silver, W. S., & Bufanio, K. M. (1996). The impact of group efficacy and group goals on group task performance. Small Group Research, 27(3), 347-359. Spink, K. S. (1990). Group cohesion and collective efficacy of volleyball teams. Journal of Sport & Exercise Psychology, 12(3), 301-311. Stajkovic, A. D., Lee, D., & Nyberg, A. J. (2009). Collective efficacy, group potency, and

779

group performance: Meta-analyses of their relationships, and test of a mediation

780

model. Journal of Applied Psychology, 94(3), 814-828. doi: 10.1037/a0015659

781

Stajkovic, A. D., & Luthans, F. (1998). Self-efficacy and work-related performance: A meta-

782

analysis. Psychological Bulletin, 124(2), 240-261. doi: 10.1037/0033-2909.124.2.240

783

Stanimirovic, R., & Hanrahan, S. J. (2004). Efficacy, affect, and teams: Is momentum a

784

misnomer? International Journal of Sport and Exercise Psychology, 2(1), 43-62. doi:

785

10.1080/1612197x.2004.9671732

786

Tasa, K., Taggar, S., & Seijts, G. H. (2007). The development of collective efficacy in teams:

787

A multilevel and longitudinal perspective. Journal of Applied Psychology, 92(1), 17-

788

27. doi: 10.1037/0021-9010.92.1.17

789

Vargas-Tonsing, T. M., & Bartholomew, J. B. (2006). An exploratory study of the effects of

790

pregame speeches on team efficacy beliefs. Journal of Applied Social Psychology,

791

36(4), 918-933.

792 793

Vargas-Tonsing, T. M., Myers, N. D., & Feltz, D. L. (2004). Coaches' and athletes' perceptions of efficacy-enhancing techniques. The Sport Psychologist, 18(4), 397-414.

794

Watson, C. B., Chemers, M. M., & Preiser, N. (2001). Collective efficacy: A multilevel

795

analysis. Personality and Social Psychology Bulletin, 27(8), 1057-1068.

RELATION BETWEEN TEAM CONFIDENCE AND PERFORMANCE 796 797

Zaccaro, S. J., Rittman, A. L., & Marks, M. A. (2001). Team leadership. Leadership Quarterly, 12(4), 451-483.

36