The influence of trained peer tutoring on tutors' motivation and performance in a French boxing setting

The influence of trained peer tutoring on tutors’motivation and performance in a French boxing setting

PASCAL LEGRAIN,1* FABIENNE D’ARRIPE-LONGUEVILLE2 and

CHRISTOPHE GERNIGON3

1Universite de Paris X-Nanterre, Paris, 2Univerisite de Sophie-Antipolis, Nice and 3Universite de Montpellier I,

Montpellier, France

Accepted 3 February 2003

The aim of this study was to examine the potential motivational and behavioural benefits of two peer tutoring

programmes for tutors in a sport setting. Differences between the sexes were also explored. Thirty two college-

age males and females, all novices on a French boxing task, were assigned to a 262 [sex6training type: physical

practice associated with trained peer tutoring (TPT) vs physical practice associated with untrained peer tutoring

(UPT)] factorial design. All participants were given six French boxing lessons of 2 h each. The TPT programme

included structured methods to prepare the participants to fulfil their role of tutors, whereas the UPT

programme did not. The results demonstrated that the TPT programme resulted in higher scores for coaching

skills. Furthermore, interaction effects revealed that the TPT programme yielded better offensive outcomes for

males and better defensive outcomes for females. Although the UPT participants expressed higher self-efficacy,

no differences emerged for intrinsic motivation and causal attributions. Finally, male tutors displayed higher

self-efficacy and offensive outcomes than female tutors. The results are discussed in the light of previous

findings in the educational and sport psychology literature.

Keywords: causal attribution, intrinsic motivation, peer tutoring, self-efficacy, sex.

Introduction

Different peer-assisted learning procedures have been

considered to be effective instructional strategies both

among academics (e.g. Topping and Ehly, 1998) and in

sport and physical activity contexts (e.g. Siedentop and

Tannehill, 2000; d’Arripe-Longueville et al., 2002).

These methods stem from two fundamental approaches:

collaborative learning (Slavin, 1990) and peer tutoring

(Sharpley and Sharpley, 1981; Topping, 1995). Peer

tutoring, the most widely known peer-assisted learning

method, is characterized by specific role-taking; at any

point, one of the peers plays the role of tutor, whereas

the other takes the role of tutee (Topping and Ehly,

1998). Different types of peer tutoring – such as tradi-

tional one-to-one peer tutoring (Cohen et al., 1982),

classwide peer tutoring (Fuchs et al., 1997) and

reciprocal peer tutoring, in which the students function

alternately as both tutor and tutee (Fantuzzo et al.,

1992) – have been advocated to favour cognitive learning

among regular and special school learners. Coaches and

physical education teachers often provide the opportu-

nity for athletes or students to learn by interacting with

each other, and peer tutoring has recently attracted the

attention of researchers in adapted (e.g. Houston-

Wilson et al., 1997) and traditional (e.g. Johnson and

Ward, 2001) physical education.

Reviews of the literature covering both research and

educational practice since the 1960s have confirmed

that peer tutoring has cognitive and social benefits for

tutors, tutees or both (e.g. Gartner et al., 1971;

O’Donnell and King, 1999). Many studies have

stressed that sometimes tutors even display more

benefits than the students they teach (Annis, 1983;

Greenwood et al., 1988; O’Donnell and O’Kelly,

1994). These studies suggest that peer tutoring might

enhance cognitive processing in the tutor by increasing

attention to and motivation for the task, and requiring a

review of existing knowledge and skills. Thus, in

addition to immediate cognitive gains, improved reten-

tion (Webb and Farivar, 1994), greater metacognitive

awareness, and better application of knowledge and

skills to new situations (Bargh and Schul, 1980;

Delquadri et al., 1983) have been reported.

* Address all correspondence to Pascal Legrain, UFRAPS, ParisX-Nanterre, 200 Avenue de la Republique, 92001 Nanterre, France.e-mail: [email protected].

Journal of Sports Sciences, 2003, 21, 539–550

Journal of Sports Sciences ISSN 0264-0414 print/ISSN 1466-447X online # 2003 Taylor & Francis LtdDOI: 10.1080/0264041031000101872

Furthermore, motivational and attitudinal benefits of

peer tutoring have been reported. For instance, Ben-

ware and Deci (1984) found that a learn-to-teach group

perceived its experience as more active and interesting,

and performed better on higher-order conceptual

understanding, than a learn-only group. Other studies

have shown the favourable effects of peer tutoring on

self-esteem (Yogev and Ronen, 1982) and on self-

confidence (Fantuzzo et al., 1992). Increased empathy

with others (Cohen et al., 1982; Yogev and Ronen,

1982) and improved attitudes towards schooling

(Lazerson et al., 1988) have also been observed

frequently. Several researchers have suggested social

role theory as a conceptual framework for explaining

affective and attitudinal changes in tutors (Allen and

Feldman, 1973; Lazerson et al., 1988). Social role

might provide the behavioural pattern by which the

individual is expected to act, but also the self-percep-

tions, emotions and attitudes belonging to this beha-

viour (Fresko, 1997). Additionally, others have

suggested that the metalearning experienced by tutors

improves self-evaluations, helps them to develop con-

trol over their learning approach options and to take

responsibility for self-direction (Linder et al., 1997).

Social interactions do not have uniform effects and

many advocates of cooperative learning (Webb, 1992)

and tutoring interventions (Topping, 1988; Fuchs et al.,

1994) have described the importance of systematic

training, implementation and monitoring procedures

to ensure that the intervention is successful. According

to social role theory, the benefits of peer tutoring for

both tutor and tutee depend on whether the role of the

tutor is ascribed or achieved, and the time and effort

devoted to the role (Lazerson et al., 1988; Fresko,

1997). Based on research findings showing that

informational structures are associated with greater

intrinsic motivation, perceived competency and ex-

ploration rather than controlling or permissive struc-

tures (Koestner et al., 1984; Deci and Ryan, 1985),

several studies have suggested that informational

structures help tutors to achieve their role more than

permissive formats. For instance, Fantuzzo et al. (1992)

showed that elementary school students in mathematics

assigned to informational structured conditions re-

ported higher global self-worth, perceived scholastic

competency, self-control, positive behavioural conduct

and mathematics gains than students assigned to non-

structured conditions.

Based on these research findings and on peer

monitoring and peer assessment procedures (O’Don-

nell and Dansereau, 1992; O’Donnell and O’Kelly,

1994), peer tutoring programmes have stressed the

importance for teachers to train students in requisite

skills, such as: (a) observation, evaluation and recording

on-task behaviour; (b) distinguishing between appro-

priate and inappropriate behaviour; (c) helping students

to make connections between causes and effects; and

(d) identifying the most important information to

communicate it effectively to peers. Specifically, John-

son and Johnson (1989) suggested that tutors should

develop knowledge to: (a) identify the motives of the

tutees’ behaviour; (b) be critical of tutees’ strategies; (c)

encourage partners to participate; and (d) summarize

clear advice. To help students assume their tutee and

tutor roles, specific instruments have been created. For

instance, lecture materials with index cards, flash cards

or worksheets have been used within reciprocal peer

tutoring programmes (Fantuzzo et al., 1992).

In summary, the educational psychology literature

suggests that peer tutoring programmes might have

academic, motivational and social benefits not only for

the tutees but also, and sometimes even more so, for the

tutors. Furthermore, effective tutoring procedures are

those in which students are trained and motivated to

achieve their role and engage in high-quality interac-

tions on suitably challenging academic tasks.

For sports and physical activities, a few studies have

suggested that effective communication and collabora-

tion between peers might contribute to psychosocial

development and motor skill acquisition. Some re-

search has reported the effects of training students to

assess each other’s performances in physical education

(Crouch et al., 1997; Ward et al., 1998). Others (Ward

and Ward, 1996; Johnson and Ward, 2001) have used

Classwide Peer Tutoring intervention in Physical

Education (CWPT-PE) to help students learn new

skills. For instance, Johnson and Ward (2001) experi-

mented with a CWPT procedure in a 20-lesson striking

unit which was made up of the following six elements:

(a) division of the class into teams of four students each;

(b) formation of peer dyads in each team; (c) use of task

cards and limited time for practice; (d) reciprocal

observation and checking; (e) posting team scores on

the number of correct trials; and (f) goal-setting for

each group at the start of each lesson. The results

indicated that during the procedure the children

performed fewer total trials, generally more correct

trials and had a higher percentage of correct trials than

at baseline. Moreover, students accurately assessed

each other’s performance more than 90% of the time.

The CWPT procedures are based on reciprocal peer

tutoring in which the students alternate between the

roles of tutor and tutee. Therefore, specific conse-

quences of being a tutor and a tutee are confounded.

Some recent studies specifically focused on the benefits

tutors can receive from peer tutoring. On the one hand,

two recent studies by Campbell and Gross (1999a,b)

reported that cross-age tutoring programmes did not

significantly alter the tutors’ attitudes towards physical

activity or their participation motivation (Campbell and

540 Legrain et al.

Gross, 1999a), and did not influence their leadership or

their self-esteem (Campbell and Gross, 1999b). On the

other hand, Legrain et al. (2003) showed the favourable

effects of an untrained peer tutoring programme

associated with physical practice, compared with

physical practice alone, in a French boxing task.

Specifically, the peer tutoring programme entailed

higher scores on boxing performance form, self-

efficacy, interest–enjoyment and personally controllable

causal attributions, and lower scores on tension–

pressure. However, no differences emerged on the

boxing performance outcomes. The lack of a significant

main effect for type of training might have been due to

the tutors not being trained to fulfil their role.

Several studies have examined the effects of un-

trained and trained tutors on the motor performance of

students with mixed results. For instance, Houston-

Wilson et al. (1997) reported that trained peer tutors

were effective at assisting students with developmental

disabilities to improve their motor performance in

physical education classes. In contrast, Webster

(1987) reported similar effects for mentally handi-

capped tutees’ performances in striking and throwing

tasks regardless of whether the tutor was trained or

untrained. According to Johnson and Ward’s (2001)

recommendations, it is unlikely that the results of

existing tutoring studies in adapted physical education

can be generalized to school physical education settings

because of the nature of both the setting and the

content. Similarly, the results of academic studies,

which mainly focused on elementary school children,

are not necessarily applicable to sports settings and to

older learners. Furthermore, existing studies in the

sport psychology literature examined the effects of

untrained and trained tutors on the performance of

tutees, but they did not consider the psychosocial and

behavioural consequences of peer tutoring for tutors

themselves.

The aim of the present study was to extend our

existing knowledge about peer tutoring in intact sport

or physical education settings. We also wished to follow

up the study by Legrain et al. (2003) by comparing the

influence of trained versus untrained peer tutoring on

the psychosocial and performance outcomes of tutors

themselves.

Based on educational studies on peer tutoring, we

made the following predictions. Because structured

interventions are supposed to help tutors in their roles

and to develop social responsibility (Lazerson et al.,

1988; Fresko, 1997), together with previous findings

among academics, trained peer tutoring was expected

to entail higher self-efficacy and intrinsic motivation

than untrained peer tutoring (Koestner et al., 1984;

Fantuzzo et al., 1992). In addition, because of greater

metacognitive awareness (Bargh and Schul, 1980;

Linder et al., 1997), trained tutors were expected to

display more appropriate coaching skills in terms of

error identification and relevance of advice than their

untrained counterparts. Because trained tutors were

supposed to develop better personal frames of refer-

ence and control over their learning approach options

(Linder et al., 1997), we expected them to demon-

strate more internal, stable and controllable causal

attributions than untrained tutors. Furthermore, ac-

cording to previous results in the physical activity

domain and the cognitive theory of motor skill

acquisition (Schmidt, 1988), structured tutoring

should help students to be more aware of their own

learning strategies and thus entail higher performance

and form scores (Johnson and Ward, 2001; Legrain et

al., 2003). Differences between the sexes were also

explored. Because French boxing is a male-oriented

physical activity (Koivula, 1995), males’ self-efficacy

beliefs (Lirgg et al., 1996; Legrain et al., 2003) and,

consequently, performance outcomes (Bandura, 1997;

Legrain et al., 2003) were expected to be higher than

those of females. Given the inconsistent findings in the

literature related to sex differences in intrinsic motiva-

tion (Amorose and Horn, 2000) and in tutoring skills

(Topping and Whiteley, 1993), no hypotheses were

forwarded for differences in intrinsic motivation and in

coachings skills between males and females. Conse-

quently, no sex6training type interaction effects were

expected.

Methods

Participants and design

Thirty-two French undergraduate physical education

students (16 males, 16 females) aged 20.1+1.3 years

(mean+s) volunteered to participate in the study. They

provided written consent before taking part. They were

selected from among 132 students from the same

university. The participants were predominantly white

and most were from a middle-class background. Young

adults were chosen because tutoring-guidance in

French boxing requires some cognitive maturity and

communication skills (Chi, 1996). Novice students

were selected for two reasons. First, because tutoring-

guidance involves many demonstrations and instruc-

tions in sport settings (d’Arripe-Longueville et al.,

2002) appear to be of particular interest in the early

phases of the acquisition of a motor skill (Fitts and

Posner, 1967; Schmidt, 1988). Second, in the educa-

tional psychology literature, students of low ability have

been shown to gain most academically from playing the

teaching role (Wiegmann et al., 1992), perhaps because

more able tutors had already acquired most of the skills

they were tutoring. Participants were told they would be

541Trained peer tutoring

filmed for the purposes of this experiment but that

confidentiality was ensured.

To determine whether the 132 students could meet

the criteria for being a study participant, they were

asked to complete questionnaires that included infor-

mation about their age, sex, number of French boxing

lessons in school and potential boxing activities outside

of school. Based on the procedure described in the

study of Legrain et al. (2003), the students took part in a

preliminary 2 h session that consisted of a French

boxing introduction. These students were then pre-

tested in a French boxing assault, in a single period of

1½ min, against an expert of the same sex. The referee

told the participants that they should try to touch their

adversary as much as possible, using fists and feet, and

to avoid being touched. Boxers were videotaped

performing the assault and were evaluated indepen-

dently by two judges familiar with boxing skills and who

were blind to the purpose of the study. The 32 selected

participants were those who the two judges agreed had

achieved a form score of 1 (single actions with a few fist

or foot combinations) and whose performance out-

comes were equal to 0.05 (see the section on

Measures). After the pre-test, the 32 participants were

assigned to one of four independent groups in a 262

[sex6training type: physical practice associated with

untrained peer tutoring (UPT) versus physical practice

associated with trained peer tutoring (TPT)] factorial

design.

Peers

Thirty-two same-sex peers with the same skill as the

participants were recruited from the same population to

act as tutees during the physical practice sessions. They

were selected according to the same pre-test procedure

as for the participants. Same-sex and same-skill peers

were chosen to maximize perceived similarity in the

dyads (Bussey and Bandura, 1984; George et al., 1992).

Task and apparatus

The task and the apparatus were similar to those used

by Legrain et al. (2003). The task in the pre-test and the

post-test consisted of a French boxing assault against a

same-sex expert. To control the participants’ perfor-

mance, the expert’s behaviour was standardized, in that

he or she could only attack with a single action and

counterattack with only one or two actions. However,

the participants were not informed of this specific

arrangement. French boxing was specifically chosen

because current competitive rules of this sport in school

require children to coach each other.

Two French boxing experts, a 21-year-old male and a

21-year-old female, were chosen to maximize perceived

similarity in the dyads in terms of age and sex. Every

participant fought against the expert in a ring (565 m)

during a single round of 1½ min. A VHS portable video

camera was used to record the participants’ behaviour

during the assault. To ensure accuracy, the operator

camera was set up for a high angle shot, 7 m from the

ring. A certified referee who was giving the instructions

and applying fundamental rules refereed the assault.

Procedure

The procedure included, in order: (a) six physical

practice sessions of 2 h each; (b) six tutoring sessions of

6 min each; (c) completion of self-efficacy question-

naires; (d) a post-test; (e) completion of intrinsic

motivation questionnaires and causal attributions ques-

tionnaires; and (f) assessment of coaching skills.

Physical practice sessions

All the participants were given six French boxing

lessons involving eight participants. The themes of the

lessons were based on the programme implemented by

Legrain et al. (2003) and taught by a certified physical

education teacher who was expert in French boxing,

and focused on fundamental technical skills. In the

two conditions, the participants practised French

boxing through learning situations and assaults with

same-sex and same-skill peers. At the end of each

lesson, the participants were invited to coach same-sex

peers.

Tutoring sessions

Each participant of the two conditions, equipped with a

towel, a sponge, a bucket and a water bottle to rinse out

the gumshield, coached peers between rounds for

6 min on six occasions during each lesson, making a

total of 36 min of coaching for the peer-tutoring

sessions. Participants in the UTP condition did not

follow any structured procedure to prepare them to

fulfil their roles as tutors. They were told to care for and

advise their peers during the rest periods, according to

the following instructions: ‘During the assault, you will

have to keep silent and observe your peer. Then, during

the rest period, you will be free to take care of your peer

with the specified equipment, and to give him/her

advice before the next round’.

Participants in the TPT group coached peers for the

same amount of time and according to the same

instructions as the UTP group, but the participants

were trained to observe and tutor peers during six

periods of 6 min each before each coaching sequence.

The total structured programme was thus 36 min long

and was made up of three parts that were based on peer

542 Legrain et al.

tutoring procedures used in cognitive learning (Johnson

and Johnson, 1989; Fantuzzo et al., 1992; O’Donnell

and Dansereau, 1992) and in physical education

settings (Johnson and Ward, 2001). Before the first

two coaching sequences, the TPT group was trained to

identify specific characteristics of correct and incorrect

behaviour (Johnson and Johnson, 1989). Participants

were invited to observe attentively a 1 min assault

between same-age novices and to assess their behaviour

using scoring sheets with pictures representing typical

standards of behaviour. Based on the Assess Pair Share

procedure of O’Donnell and Dansereau (1992), the

participants were invited to: (a) focus attention on only

one component of the boxers’ behaviour; (b) assess the

boxers’ form and performance individually; (c) com-

pare their assessments to those of another tutor; and (d)

compare their assessments to those of the teacher. Time

limits were set for each step of this procedure, which

was replicated five times, each time focusing on one of

the following technical themes: (a) distance; (b)

protection; (c) control of force; (d) coordination of

actions; and (e) variety of lines attacked.

Before the third and fourth sequences of coaching,

the tutors were trained to identify causes and con-

sequences of typical novice behaviour in French boxing.

Specifically, the tutors first had to observe an assault

between novices. While watching, they were invited to

use a checklist of typical technical difficulties in French

boxing and to identify the specific components of the

boxer’s behaviour which needed to be improved. Three

types of errors were determined: (a) strategic errors

(e.g. standing in the corner of the ring without doing

anything); (b) perceptual errors (e.g. misjudging

distance); and (c) execution errors (e.g. wayward body

movements). The TPT participants also had to assess

the boxer’s offensive and defensive performance using

scoring cards (Fantuzzo et al., 1992; Johnson and

Ward, 2001).

Finally, before the fifth and sixth sequences of

coaching, the programme aimed to help the tutors

formulate appropriate advice in French boxing. The

TPT group was invited to observe four theme assaults,

and to give informative feedback according to the

following format: (a) type of error; (b) cause of the

error; (c) required behaviour; and (d) appropriate

advice. When several errors were identified, the tutors

learned to set up a hierarchy of errors. The teacher’s

own assessments and points of view served as a

reference in this process. Rest period constraints

(1 min) had to be respected to formulate advice.

Administration of questionnaires and post-test

Before the post-test, the participants were invited to

complete self-efficacy questionnaires adapted from

Bandura and Adams (1977). The conditions of the

post-test were similar to the pre-test. After the post-test,

French versions of the Intrinsic Motivation Inventory of

McAuley et al. (1989) and of the Revised Causal

Dimension Scale of McAuley et al. (1992) were

administered.

Coaching skills assessment

Finally, the participants were invited to watch a 1½ min

videotape of a French boxing assault between two

same-age and same-sex beginners dressed in red and in

blue, respectively, to identify errors and to propose

virtual advice to one selected boxer on a specific sheet.

The following instructions were given to the partici-

pants: ‘You are going to observe an assault between two

novices in French boxing. Please pay close attention

because you will see it only once. At the end, you will

have to indicate in the first column of the table of the

sheet the errors you identified for the boxer dressed in

blue. Furthermore, you will have to indicate in the

second column as precisely as possible the advice you

should give to him/her to perform better’. At the end of

the video, the participants had 10 min to complete their

table.

Measures

Three different boxing scales created by Legrain (1998)

were used to measure performance outcomes and form

scores in the present study. These measures, usually

used by referees to award points in French boxing, were

assessed using the ratings of two independent judges,

expert in boxing, who were blind to the aims of the

study and to the experimental conditions.

Offensive performance

Consistent with Johnson and Ward (2001), offensive

performance was scored by dividing the number of

successful offensive actions (attacks and counterattacks

with foot or fist) by the total number of attempts

displayed during the assault.

Defensive performance

Participants’ defensive performance was scored by

dividing the number of effective blows administered

by the expert’s foot or fist by the total number of blows

landed by the expert in attack and counterattack.

Boxing form score

The form score was calculated according to the series of

blows which are part of the technical score in French

543Trained peer tutoring

boxing. The raters could attribute: (a) 1 point for a

single foot or fist action; (b) 2 points for a foot or fist

combination; (c) 3 points for a series of foot and fist

blows; and (d) 4 points for a series combining three

actions of feet and fists. An average score was calculated

for each participant on the basis of the total number of

technical actions displayed during the assault.

The raters practised the scoring scheme using pilot

videotapes of 15 same-age students. They clarified

discrepant coding before viewing the videotapes sepa-

rately for subsequent data analysis. On pilot videotapes,

97% agreement was obtained for performance form and

95.3% agreement was reached for performance out-

comes. Inter-rater reliabilities were calculated and are

discussed in the Results section.

Self-efficacy beliefs

Level and strength of efficacy judgements for perfor-

mance form and outcomes were measured by adapting

procedures recommended by Bandura and Adams

(1977) and used by Legrain et al. (2003). The form

scores scale included the four techniques used to rate

form scores. Because attaining a given score did not

imply attaining lower scores, and contrary to Ban-

dura’s procedure, the participants were asked to

record only the best score they expected to get by

the end of the assault. Although the participants did

not receive any training on the self-efficacy measures,

they received information about the criteria used to

evaluate their performance form and outcomes and

about their initial form score and performance out-

comes after the pre-test. We believed that this

information would help them determine their potential

scores, because the performance form score and

outcomes scales and the related self-efficacy scales

were based on the same criteria. Regarding self-

efficacy for performance outcomes, the participants

had to indicate their self-efficacy for offensive out-

comes (the number of times the participant expected

to touch the expert during the assault relative to his or

her number of attempts) and their self-efficacy for

defensive outcomes (the number of times the partici-

pant expected to be touched by the expert during the

assault relative to his or her number of attempts). Self-

efficacy for performance outcomes was obtained by

dividing self-efficacy for offensive performance by self-

efficacy for defensive performance. To measure the

strength of self-efficacy, the participants were invited

to indicate their confidence in reaching the expected

score on a scale ranging from 10% (not sure) to 100%

(totally sure). The strength of self-efficacy for perfor-

mance outcomes was calculated as the mean of the

strength of self-efficacy for offensive outcomes and of

the strength of self-efficacy for defensive outcomes.

Intrinsic motivation

We used the French version of the Intrinsic Motiva-

tion Inventory (IMI) of McAuley et al. (1989)

adapted to French boxing by Legrain et al. (2003).

The IMI consists of 18 items scored on a Likert scale

of 1 (strongly disagree) to 7 (strongly agree). Four

subscales are distinguished (interest–enjoyment, per-

ceived competence, effort–importance and tension–

pressure) with a minimum of four items per subscale.

According to previous factor analyses conducted on

similar participants involved in a French boxing task

(Legrain et al., 2003), five items (4, 7, 13, 15, 16) of

the original IMI scale were not employed on the

French IMI scale in the present study. The following

factors, which were conceptually consistent with the

original IMI scale and which had demonstrated good

internal consistency, were retained. The first factor

consisted of three items (3, 6, 12) reflecting effort–

importance on the original scale; the second factor

consisted of three items (5, 10, 11) reflecting tension–

pressure on the original scale; the third factor

consisted of three items (1, 8, 17) reflecting

interest–enjoyment on the original scale; and the

fourth factor consisted of three items (2, 9, 14)

reflecting perceived competence on the original scale.

In the present study, the reliability coefficients

(Cronbach’s alpha) for these four factors were 0.78,

0.77, 0.74 and 0.83, respectively, thus indicating

satisfactory internal consistency.

Causal attributions

The French version of the Revised Causal Dimension

Scale (CDSII) of McAuley et al. (1992), used by

Legrain et al. (2003), was used to assess causal

attributions. The CDSII allowed participants: (a) to

indicate how well they thought they had performed

on a 9-point Likert scale; (b) to provide their own

open-ended attribution for their performance out-

come; and (c) to code that attribution along causal

dimensions. Twelve semantic differential scales, with

three items representing each of the dimensions

(locus of causality, stability, personal control and

external control), make up the CDSII. Values can

range from 3 to 27, with higher values representing

attributions that are more internal, stable and either

personally or externally controllable. According to

previous factor analyses conducted on similar partici-

pants involved in a French boxing task (Legrain et al.,

2003), two factors on the original scale were retained

to make up the French CDSII scale in the present

study. The first factor consisted of three items (2, 4,

10) reflecting personal control on the original scale;

the second factor consisted of three items (1, 6, 9)

544 Legrain et al.

reflecting locus of causality on the original scale. In

the present study, the reliability coefficients (Cron-

bach’s alpha) for these two factors were 0.79 and

0.69, respectively, thus indicating satisfactory internal

consistency.

Coaching skills

Two experts in French boxing who were blind to the

study conditions watched the same French boxing

assault as the participants. They independently

identified five major, three secondary and two minor

technical errors and formulated appropriate advice.

For each participant, a 20 point score related to the

identification of errors and a 20 point score related to

the appropriateness of advice were calculated, accord-

ing to the experts’ assessments. According to the

experts’ assessments, the following criteria were used:

(a) 4 points for identifying a major error or

formulating important advice; (b) 3 points for

identifying a secondary error or formulating second-

ary advice; (c) 2 points for identifying a minor error

or formulating minor advice; and (d) 1 point for

identifying an irrelevant error or formulating irrele-

vant advice (not identified by the two experts). The

final score for coaching skills was the sum of the two

independent scores related to the identification of

errors and to the appropriateness of advice. Inter-rater

reliabilities for these judgements yielded a high

percentage of agreement both for the scores related

to identifying errors (98.2%) and for advice scores

(97.3%).

Results

Reliability analyses

Inter-rater reliabilities for performance judgements at

post-test were assessed by calculating the percentage

of agreement between the two judges and the

intraclass correlation coefficient (R) using a two-way

analysis of variance ANOVA (participants6raters)

with repeated measures on the last factor. These

analyses yielded a high percentage of agreement for

performance form (90.6%), offensive performance

(92.8%) and defensive performance (93.8%), and

good intraclass correlation coefficients (R=0.98,

R=0.97 and R=0.98, respectively). Therefore, the

judges’ ratings of performance were averaged. In

addition, a videotape of eight students, two from each

group, was developed for calculating intra-rater

reliability. The performance form and outcomes of

the eight participants were recoded 8 months later,

and both coders replicated their scores 95.2% of the

time.

Training type and sex differences on study

variables

Means and standard deviations for all measures are

presented in Table 1. Four separate 262 (training

type6sex) multivariate analyses of variance were

performed to examine training type and sex differences

on: (a) offensive performance, defensive performance,

form score and coaching skills; (b) self-efficacy beliefs;

(c) intrinsic motivation; and (d) causal attributions.

Follow-up tests were conducted with univariate ana-

lyses of variance. Bonferroni’s corrections were used to

maintain an overall alpha of 0.05 in univariate F-tests,

and thus the alpha level necessary to demonstrate

significance was P=0.013. Effect sizes (ES) were also

calculated using pooled standard deviations, based on

the formula ES= (mean 17mean 2)/s (Hedges and

Olkin, 1985).

Behavioural variables

The results of the multivariate analysis of variance on

offensive performance, defensive performance, form

score and coaching skills indicated a significant main

effect of training type (Wilks’ l=0.36, F4,25 = 11.20,

P50.001) and a main effect of sex (Wilks’ l=0.53,

F4,25 = 5.52, P50.002). The interaction between train-

ing type and sex was also significant (Wilks’ l=0.49,

F4,25 = 6.60, P50.001).

Offensive performance

For offensive performance, analysis of variance

indicated a main effect of training type (F1,28 = 23.5,

P50.001, ES=1.17) and of sex (F1,28 = 18.3,

P50.001, ES=1.05). There was also a significant

interaction between training type and sex (F1,28 = 22.2,

P50.001). Tests of the significant interaction indicated

that male TPT participants demonstrated significantly

(P50.001) higher offensive performance (mean=

0.50+0.08) than male UPT participants (mean=

0.25+0.09) (ES=2.94), and also the females in both

the TPT and UPT conditions (mean=0.26+0.08 and

0.26+0.05; ES=3.00 and 3.60, respectively).

Defensive performance

For defensive performance, analysis of variance indi-

cated a main effect of training type (F1,28 = 13.4,

P50.001, ES=1.11). The main effect of sex was not

significant (F1,28 = 0.35, P=0.56). However, the inter-

action between training type and sex was significant

(F1,28 = 11.8, P50.002). Paired comparisons indicated

that female TPT participants demonstrated signifi-

cantly (P50.001) better defensive performance

545Trained peer tutoring

(mean=0.33+0.06) than female UPT participants

(mean=0.53+0.05; ES=3.62).

Boxing form score

For form score, the analysis of variance indicated that

the main effect of training type (F1,28 = 0.62, P=0.44),

the main effect of sex (F1,28 = 0.20, P=0.66) and the

interaction between the two (F1,28 = 0.05, P=0.82)

were not significant.

Coaching skills

For coaching skills, analysis of variance revealed only a

significant main effect of training type (F1,28 = 17.0,

P50.001, ES=1.49). The TPT group

(mean=22.50+5.29) scored higher than the UPT

group (mean=13.88+6.25).

Self-efficacy beliefs

A 262 multivariate analysis of variance (training

type6sex) was computed on the level and the strength

of self-efficacy judgements. The results indicated a

significant main effect of training type (Wilks’ l=0.66,

F4,25 = 3.21, P=0.03) and a main effect of sex (Wilks’

l=0.52, F4,25 = 5.71, P50.002). However, the inter-

action between training type and sex (Wilks’ l=0.81,

F4,25 = 1.44, P=0.25) was not significant.

For level of self-efficacy for performance, analysis of

variance indicated no main effect of training type

(F1,28 = 0.01, P=0.91) or main effect of sex

(F1,28 = 0.95, P=0.34). For strength of self-efficacy

for performance, analysis of variance yielded a sig-

nificant main effect of sex (F1,28 = 8.17, P=0.008,

ES=0.94). Males scored higher (mean=

62.81+13.41) than females (mean=52.66+7.33).

The main effect for training type was not significant

(F1,28 = 5.85, P50.02). For level of self-efficacy for

form score, analysis of variance yielded a significant

main effect of training type (F1,28 = 7.30, P=0.012,

ES=0.88) and of sex (F1,28 = 7.30, P=0.012,

ES=0.88). Participants in the UPT condition displayed

higher scores (mean=2.69+0.40) than participants in

the TPT condition (mean=2.25+0.58). In addition,

males had higher scores (mean=2.69+0.60) than

females (mean=2.25+0.37). For strength of self-

efficacy for form score, the main effects for training

type (F1,28 = 1.26, P=0.27) and sex (F1,28 = 2.16,

P=0.15) were not significant.

Table 1. Results by training type and sex (n=32; mean+s)

Training type (n=16) Sex (n=16)

UPT TPT Males Females

Behavioural variables

Offensive performance 0.26+0.07 0.38+0.14 0.38+0.15 0.26+0.06

Defensive performance 0.47+0.09 0.37+0.09 0.41+0.10 0.43+0.12

Boxing form score 1.38+0.17 1.44+0.27 1.39+0.22 1.43+0.24

Coaching skills 13.88+6.25 22.50+5.29 17.50+8.08 18.88+6.35

Self-efficacy beliefs

Level of self-efficacy for performance 0.52+0.17 0.53+0.23 0.56+0.26 0.49+0.11

Strength of self-efficacy for performance 62.03+10.77 53.44+11.54 62.81+13.41 52.66+7.33

Level of self-efficacy for form scores 2.69+0.40 2.25+0.58 2.69+0.60 2.25+0.37

Strength of self-efficacy for form scores 60.63+9.29 56.56+11.93 61.25+12.58 55.94+8.00

Intrinsic motivation

Interest–enjoyment 18.69+1.85 16.81+3.62 17.19+3.35 18.31+2.55

Tension–pressure 10.94+4.07 13.63+4.81 13.00+5.53 11.56+3.46

Effort–importance 16.38+2.87 13.44+3.54 14.88+3.34 14.94+3.77

Perceived competence 12.13+2.66 11.56+3.90 11.69+4.01 12.00+2.50

Causal attributions

Personal control of causality 19.50+5.29 18.69+5.76 18.13+7.21 20.06+2.74

Internal locus of causality 19.25+3.62 18.81+5.69 17.75+6.09 20.31+2.24

Note: UPT=untrained peer tutoring; TPT= trained peer tutoring.

546 Legrain et al.

Intrinsic motivation

Results of the multivariate analysis of variance revealed

a main effect of training type (Wilks’ l=0.68,

F4,25 = 2.94, P=0.04). Neither the main effect of sex

(Wilks’ l=0.95, F4,25 = 0.36, P=0.83) nor the inter-

action between training type and sex (Wilks’ l=0.89,

F4,25 = 0.74, P=0.57) were significant. Univariate

analyses indicated no significant effect of training type

on the measures of effort–importance (F1,28 = 6.69,

P=0.015), interest–enjoyment (F1,28 = 3.32, P=0.08),

tension–pressure (F1,28 = 2.89, P50.10) or perceived

competence (F1,28 = 0.21, P=0.65).

Causal attributions

Results of the multivariate analysis indicated that the

main effects for training type (Wilks’ l=0.99,

F2,27 = 0.09, P=0.91) and sex (Wilks’ l=0.91,

F2,27 = 1.27, P=0.30) were not significant; nor was

the interaction between training type and sex (Wilks’

l=0.99, F2,27 = 0.20, P=0.82).

Discussion

The aim of the present study was to examine whether

trained peer tutoring in a French boxing setting could

contribute to higher motivational outcomes, perfor-

mance and coaching skills than untrained peer tutoring.

Differences between the sexes were also explored. The

hypothesized relationships were examined using a series

of multivariate analyses. The results revealed only

partial support for the predicted findings.

Based on social role theory and related peer tutoring

studies (Lazerson et al., 1988; Fresko, 1997), we

predicted that a structured programme of tutoring would

help the tutors fulfil their role and would thus entail

higher self-efficacy (Fantuzzo et al., 1992) and intrinsic

motivation (Koestner et al., 1984) than untrained peer

tutoring. However, these expected differences were not

observed in the present French boxing setting. Surpris-

ingly, untrained tutors demonstrated higher self-efficacy

for form score than their trained counterparts, while no

differences appeared on the other self-efficacy variables

or on intrinsic motivation.

Consistent with our predictions, however, trained

tutors demonstrated better coaching skills – that is, they

identified errors more precisely and gave more relevant

advice than untrained tutors. Together with the results

of Linder et al. (1997) for university physics tutors, it can

be assumed that the structured French boxing tutoring

programme helped the tutors develop declarative knowl-

edge and metalearning strategies that were useful to

coaching skills. These findings are consistent with those

of previous studies with academics, which reported that

peer tutoring favoured metacognitive awareness and

application of knowledge and skills to new situations

(Bargh and Schul, 1980). These results also support

physical education studies that showed that structured

peer tutoring increased the quality of peer assessment

(Crouch et al., 1997; Johnson and Ward, 2001).

Our findings also suggest that, because of better

metacognitive awareness, the trained tutors might have

increased their awareness of their own skill and of the

task difficulty (Linder et al., 1997). Indeed, discrepan-

cies between efficacy beliefs and performance are more

likely to occur when one has little information on which

to base efficacy judgements, such as when one is

learning a skill (Bandura, 1997; Moritz et al., 2000).

Thus it might be speculated that compared with the

untrained tutors, the tutors involved in a structured

peer tutoring procedure demonstrated lower but

possibly more realistic perceived competence and self-

efficacy beliefs. Finally, one possible reason why the

expected differences in intrinsic motivation did not

emerge is that the structured programme was seen by

the participants as a controlling context, which is known

to entail lower intrinsic motivation than informative

structures (Koestner et al., 1984).

Regarding causal attributions and based on the study

of Linder et al. (1997), we predicted that trained tutors

would demonstrate more internal, stable and control-

lable causal attributions than untrained tutors. Given

their lack of internal consistency, two subscales of the

original Causal Dimension Scale (external control and

stability) were removed from the analyses. Contrary to

our expectations, no differences emerged on the

measure of personal control or on the locus of causality.

Although the results observed on coaching skills suggest

that the trained tutors have developed better declarative

knowledge in French boxing than untrained tutors, this

knowledge did not appear to help them develop better

personal frames of reference and control over their own

learning approach options. It should be acknowledged

that the methods used in other studies were somewhat

different. Indeed, Linder et al. (1997) examined causal

attributions through qualitative data analysis, whereas

the CDSII was used in the present study. This may

account for why the present results did not conform to

those reported for academics.

Consistent with our predictions, however, trained

tutors obtained higher offensive and defensive scores

than untrained tutors. Interactions between sex and

training type were also observed. Trained male tutors

demonstrated higher offensive performance than their

untrained counterparts. In addition, defensive perfor-

mance scores were higher among trained female tutors

than among untrained females. These results, obtained

in a traditional one-to-one peer tutoring setting, are

547Trained peer tutoring

consistent with earlier studies of reciprocal peer

tutoring which revealed that structured peer tutoring

conditions helped develop conceptual learning (Fan-

tuzzo et al., 1992) and enhanced motor skill acquisition

(Johnson and Ward, 2001). They also complement

previous studies in the adapted physical education

literature that revealed the positive influence of trained

tutors on tutees’ performance (Houston-Wilson et al.,

1997).

Although we expected trained peer tutoring that

entailed higher coaching skills in terms of error

identification and advice to also entail higher form

score, no training type differences emerged on this

measure. The current debates in the motor learning

literature might offer some explanation to account for

this result. These debates indicate that, contrary to

conceptual learning settings, metacognitive awareness

might not necessarily be required in a sports setting, as

evidenced for relatively brief, complex, self-paced

repetitive acts (Singer et al., 1993) and open motor

skills (Magill, 1998). The fact that trained tutors

demonstrated better coaching skills than untrained

tutors but similar form scores would thus support the

equivocal role of instructions and explicit knowledge

during the first stages of motor skill acquisition (Wulf

and Weigelt, 1997).

The design of the studies, and specifically the type of

peer tutoring and the characteristics of the participants

involved, may account for some of the inconsistencies.

For instance, although a one-to-one peer tutoring

procedure with fixed tutor and tutee roles was used in

the present research, many previous studies referred to

reciprocal peer tutoring in which the students alternated

between the tutor and the tutee roles (Fantuzzo et al.,

1992; Johnson and Ward, 2001). Furthermore, the

participants in these studies were elementary school

students, while undergraduate physical education stu-

dents were involved in the present investigation.

Based on previous findings in the sport psychology

literature (Lirgg et al., 1996; Legrain et al., 2003), we

predicted that males’ self-efficacy beliefs, form score

and performance would be higher than those of

females. Consistent with our hypotheses, females

expressed lower self-efficacy for form score and less

certain judgements about their performance than their

male counterparts. Males also demonstrated a better

offensive performance than females. This pattern of

results would confirm the masculine connotation of

the tasks (Lirgg et al., 1996). As a masculine-oriented

activity (Koivula, 1995), the French boxing task might

have appeared less familiar to females than to males,

thus rendering the latter more confident (Legrain et

al., 2003) and offensive. However, the fact that no sex

differences were noted for intrinsic motivation, causal

attributions, form scores or defensive performance

would suggest that both females and males viewed the

task as gender-appropriate (Morgan et al., 1996).

These conflicting results suggest further research is

required.

There are some limitations to our study that should

be addressed. First, the sample size and the absence of a

control group suggest caution in drawing conclusions.

Second, the use of single items to measure self-efficacy

and the psychometric properties of the French versions

of the IMI and the CDSII used in the present study also

limit the generalizability of the findings. Third, the age

of the participants and the type of one-to-one peer

tutoring with fixed roles adopted in this study might

suffer from a lack of ecological validity. Many peer

tutoring studies have focused on elementary/primary

school settings and most practitioners give students the

opportunity to alternate between tutee and tutor roles

and thus refer to reciprocal peer tutoring procedures.

Therefore, our design limits generalizability to the

practitioner’s world.

In conclusion, the present study provides interesting

findings on peer tutoring sports. First, consistent with

earlier educational studies, we have revealed that novice

students acting as trained tutors in a French boxing

setting demonstrated better coaching skills than un-

trained tutors. Furthermore, specific benefits of trained

peer tutoring on males’ offensive performance and on

females’ defensive outcomes were noted. From a

practical point of view, this would mean that trained

tutors could both better advise peers and enhance their

own performance than untrained tutors. The present

findings could therefore help coaches and physical

education teachers to reconsider their role in terms of

preparation and management of cooperative small

groups. However, the fact that untrained tutors

expressed higher self-efficacy, whereas similar effects

were observed on intrinsic motivation, causal attribu-

tions and form scores regardless of whether the tutors

were trained or untrained, calls for caution and further

research in this area.

Future studies might address the influence of trained

tutoring in tasks more tailored to the participants’ level

of skill, and might measure the perceived difficulty of

the task and the perception of the learning context,

which are known to influence self-efficacy (Bandura,

1997) and intrinsic motivation (Koestner et al., 1984).

Research should also examine the influence of peer

tutoring strategies across various physical activity

contexts. Furthermore, intervention approaches might

require tailoring to the particular personal character-

istics of peers. Because peer tutoring has been shown to

require some cognitive maturity (Chi, 1996), young

children might need specific training. The present study

also highlights a need for further research on sex

differences in peer tutoring.

548 Legrain et al.

Finally, there is a need to identify under what

conditions structuring interactions would be productive

in the physical domain. For that purpose, examining the

relationship between peer tutoring and self-regulated

learning strategies (Zimmerman and Kitsantas, 1997)

might be appropriate. These theoretically based ap-

proaches should both extend our knowledge of the

process and outcomes of peer-assisted learning in the

physical activity domain, and provide coaches and

physical education teachers with additional ways to

foster positive peer interactions.

Acknowledgements

The authors thank the students of the participating schools.

Sincere appreciation is also extended to Vincent Dru for

methodological assistance and Stephanie Hanrahan for help-

ful comments. A part of this paper was presented at the

International Congress of the French Society of Sport

Psychology, Paris, July 2000.

References

Allen, V.L. and Feldman, R.S. (1973). Learning through

tutoring: low achieving children as tutors. Journal of

Experimental Education, 42, 1–5.

Amorose, A.J. and Horn, T.S. (2000). Intrinsic motivation:

relationships with collegiate athletes’ gender, scholarship

status, and perceptions of their coaches’ behavior. Journal

of Sport and Exercise Psychology, 22, 63–84.

Annis, L.F. (1983). The process and effects of peer tutoring.

Human Learning, 2, 39–47.

Bandura, A. (1997). Self-efficacy: The Exercise of Control. New

York: Freeman.

Bandura, A. and Adams, N.E. (1977). Analysis of self-

efficacy theory of behavioural change. Cognitive Therapy

and Research, 1, 287–308.

Bargh, J.A. and Schul, Y. (1980). On the cognitive benefits

of teaching. Journal of Educational Psychology, 72, 593–604.

Benware, C.A. and Deci, E.L. (1984). Quality for learning

with an active versus passive motivational set. American

Educational Research Journal, 21, 755–765.

Bussey, K. and Bandura, A. (1984). Influence of gender

constancy and social power on sex linked modeling.

Journal of Personality and Social Psychology, 47, 1292–1302.

Campbell, H. and Gross, J. (1999a). Cross-age tutoring:

effects on children’s participation motivations and atti-

tudes toward physical activity. Communication to the

Annual Conference of the Association for the Advancement of

Applied Sport Psychology, Banff, Alberta, Canada.

Campbell, H. and Gross, J. (1999b). Cross-age tutoring: an

effective leadership alternative for sports skills. Commu-

nication to the Annual Conference of the Association for the

Advancement of Applied Sport Psychology, Banff, Alberta,

Canada.

Chi, M.T.H. (1996). Constructing self-explanations and

scaffolded explanations in tutoring. Applied Cognitive

Psychology, 10, 33–49.

Cohen, P.A., Kulik, J.A. and Kulik, C.C. (1982). Educa-

tional outcomes of peer tutoring: a meta-analysis of

findings. American Educational Research Journal, 19, 237–

248.

Crouch, D.W., Ward, P. and Patrick, C.A. (1997). The

effects of peer-mediated accountability on task accom-

plishment during volleyball drills in elementary physical

education. Journal of Teaching in Physical Education, 17,

26–39.

d’Arripe-Longueville, F., Huet, M.L., Gernigon, C., Win-

nykamen, F. and Cadopi, M. (2002). Peer assisted

learning in the physical activity domain: dyad type and

gender differences. Journal of Sport and Exercise Psychology,

24, 219–238.

Deci, E.L. and Ryan, R.M. (1985). Intrinsic Motivation and

Self-determination in Human Behavior. New York: Plenum

Press.

Delquadri, J., Greenwood, C.R., Stretton, K. and Hall, R.V.

(1983). The peer tutoring game: a classroom procedure

for increasing opportunity to respond and spelling

performance. Education and Treatment of Children, 6,

225–239.

Fantuzzo, J.W., King, A.K. and Heller, L.R. (1992). Effects

of reciprocal peer tutoring on mathematics and school

adjustment: a component analysis. Journal of Educational

Psychology, 3, 331–339.

Fitts, P.M. and Posner, M.I. (1967). Human Performance.

Belmont, CA: Brooks/Cole.

Fresko, B. (1997). Attitudinal change among university

student tutors. Journal of Applied Social Psychology, 27,

1277–1301.

Fuchs, D., Fuchs, L.S., Mathes, P.G. and Simmons, D.C.

(1997). Peer-assisted learning strategies: making class-

rooms more responsive to diversity. American Educational

Research Journal, 34, 174–206.

Fuchs, L.S., Fuchs, D., Bentz, J., Philips, N.B. and Hamlett,

C.L. (1994). The nature of student interactions during

peer tutoring with and without prior training and

experience. American Educational Research Journal, 31,

75–103.

Gartner, A., Kohler, M.C. and Riessman, F. (1971).

Children Teach Children: Learning by Teaching. New York:

Harper & Row.

George, T.R., Feltz, D.L. and Chase, M.A. (1992). Effects

of model similarity on self-efficacy and muscular endur-

ance: a second look. Journal of Sport and Exercise

Psychology, 14, 237–248.

Greenwood, C.R., Carta, J.C. and Hall, R.V. (1988). The

use of peer tutoring strategies in classroom management

and educational instruction. School Psychology Review, 17,

258–275.

Hedges, L.V. and Olkin, I. (1985). Statistical Methods for

Meta-Analysis. Orlando, FL: Academic Press.

Houston-Wilson, C., Dunn, J., van der Mars, H. and

McCubbin, J. (1997). The effect of peer tutors on motor

performance in integrated physical education classes.

Adapted Physical Activity Quarterly, 14, 298–313.

549Trained peer tutoring

Johnson, D.W. and Johnson, R.T. (1989). Cooperation and

Competition: Theory and Research. Edina, MN: Interaction

Book Co.

Johnson, M. and Ward, P. (2001). Effects of classwide peer

tutoring on correct performance of striking skills in 3rd

grade physical education. Journal of Teaching Physical

Education, 20, 247–263.

Koestner, R., Ryan, R.M., Bernieri, F. and Holt, K. (1984).

Setting limits on children’s behavior: the differential effects

of controlling vs. informational styles on intrinsic motiva-

tion and creativity. Journal of Personality, 52, 233–248.

Koivula, N. (1995). Ratings of gender appropriateness of

sports participation: effects of gender-based schematic

processing. Sex Roles, 33, 543–558.

Lazerson, D.B., Foster, H.L., Brown, S.I. and Hummel,

J.W. (1988). The effectiveness of cross-age tutoring with

truant junior high students with learning disabilities.

Journal of Learning Disabilities, 21, 253–255.

Legrain, P. (1998). Boxe Francaise. Apprentissage et Enseigne-

ment. Paris: Editions revue EPS, Les Cahiers du Sport.

Legrain, P., d’Arripe-Longueville, F. and Gernigon, C.

(2003). Peer tutoring in a sport setting: are there any

benefits for tutors? The Sport Psychologist, 17, 77–94.

Linder, C.J., Leonard-McIntyre, C., Marshall, D. and

Nchodu, M.R. (1997). Physics tutors’ metalearning devel-

opment through an extension of Schon reflective practice.

International Journal of Science Education, 19, 821–833.

Lirgg, C.D., George, T.R., Chase, M.A. and Ferguson,

R.H. (1996). Impact of conception of ability and sex-type

of task on male and female self-efficacy. Journal of Sport

and Exercise Psychology, 18, 426–434.

Magill, R.A. (1998). Motor Learning: Concepts and Applica-

tions. New York: McGraw-Hill.

McAuley, E., Duncan, T.E. and Tammen, V.V. (1989).

Psychometric properties of the Intrinsic Motivation Inven-

tory in a competitive sport setting: a confirmatory factor

analysis. Research Quarterly for Exercise and Sport, 60, 48–58.

McAuley, E., Duncan, T.E. and Russell, D. (1992).

Measuring causal attributions: the revised Causal Dimen-

sion Scale (CDS II). Personality and Social Psychology

Bulletin, 18, 566–573.

Morgan, L.K., Griffin, J. and Heyward, V.H. (1996).

Ethnicity, gender, and experience effects on attributional

dimensions. The Sport Psychologist, 10, 4–16.

Moritz, S.E., Feltz, D.L., Fahrbach, K.R. and Mack, D.E.

(2000). The relation of self-efficacy measures to sport

performance: a meta-analytic review. Research Quarterly for

Exercise and Sport, 71, 280–294.

O’Donnell, A.M. and Dansereau, D.F. (1992). Scripted co-

operation in student dyads: a method for analysing and

enhancing academic learning and performance. In Interac-

tion in Co-operative Groups: The Theoretical Anatomy of Group

Learning (edited by R. Hertz-Lazarowitz and N. Miller), pp.

120–141. Cambridge: Cambridge University Press.

O’Donnell, A.M. and King, A. (eds) (1999). Cognitive

Perspectives on Peer Learning. Mahwah, NJ: Lawrence

Erlbaum Associates.

O’Donnell, A.M. and O’Kelly, J. (1994). Learning from

peers: beyond the rhetoric of positive results. Educational

Psychology Review, 6, 321–349.

Schmidt, R.A. (1988). A schema theory of discrete motor

skill learning. Psychological Review, 82, 225–250.

Sharpley, A.M. and Sharpley, C.F. (1981). Peer tutoring: a

review of literature. Collected Original Resources in Educa-

tion, 5, 7–11.

Siedentop, D. and Tannehill, D. (2000). Developing Teaching

Skills in Physical Education, 4th edn. Palo Alto, CA:

Mayfield.

Singer, R.N., Lidor, R. and Cauraugh, J.H. (1993). To be

aware or not aware? What to think about while learning

and performing a motor skill. The Sport Psychologist, 7, 19–

30.

Slavin, R.E. (1990). General education under the regular

education initiative: how must it change? Remedial and

Special Education, 11, 40–50.

Topping, K.J. (1988). The Peer Tutoring Handbook: Promoting

Co-operative Learning. London: Croom Helm.

Topping, K.J. (1995). Paired Reading, Spelling and Writing:

The Handbook for Teachers and Parents. London: Cassell.

Topping, K. and Ehly, S. (1998). Peer-Assisted Learning.

Mahwah, NJ: Lawrence Erlbaum Associates.

Topping, K. and Whiteley, M. (1993). Sex diferences in the

effectiveness of peer tutoring. School Psychology Interna-

tional, 14, 57–67.

Ward, P. and Ward, M.C. (1996). The effects of classwide

peer tutoring on correct cardiopulmonary resuscitation

performance by physical education majors. Journal of

Behavioral Education, 6, 331–342.

Ward, P., Smith, S.L., Makasci, K. and Crouch, D. (1998).

Differential effects of peer-mediated accountability in

elementary physical education majors. Journal of Teaching

in Physical Education, 17, 442–452.

Webb, N.M. (1992). Testing a theoretical model of student

interaction and learning in small groups. In Interaction in

Cooperative Groups: The Theoretical Anatomy of Group

Learning (edited by R. Hertz-Lazarowitz and N. Miller),

pp. 102–119. New York: Cambridge University Press.

Webb, N. and Farivar, S. (1994). Promoting helping

behavior in cooperative small groups in middle school

mathematics. American Educational Research Journal, 31,

369–395.

Webster, G.E. (1987). Influence of peer tutors upon

academic learning time – physical education of mentally

handicapped students. Journal of Teaching in Physical

Education, 7, 393–403.

Wiegmann, D.A., Dansereau, D.F. and Patterson, M.E.

(1992). Co-operative learning: effects of role-playing and

ability on performance. Journal of Experimental Education,

60, 109–116.

Wulf, G. and Weigelt, C. (1997). Instructions about physical

principles in learning a complex motor skill: to tell or not

to tell. Research Quarterly for Exercise and Sport, 68, 362–

367.

Yogev, A. and Ronen, R. (1982). Cross-age tutoring: effects

on tutors’ attributes. Journal of Educational Research, 75,

261–268.

Zimmerman, B.J. and Kitsantas, A. (1997). Developmental

phases in self-regulation: shifting from process goals to

outcome goals. Journal of Educational Psychology, 89, 29–

36.

550 Legrain et al.