Preschool-aged children’s jumps: Imitation performances

Journal of Electromyography and Kinesiology 20 (2010) 322–329

Contents lists available at ScienceDirect

Journal of Electromyography and Kinesiology

journal homepage: www.elsevier .com/locate / je lek in

Preschool-aged children’s jumps: Imitation performances

Lazhar Labiadh, Marie-Martine Ramanantsoa, Eveline Golomer *

Eveline Golomer, Laboratoire EA 4070, Equipe Action Mouvement Adaptation–UFR STAPS 1 rue Lacretelle, 75015 Paris, France

a r t i c l e i n f o

Article history:Received 9 November 2007Received in revised form 29 April 2009Accepted 20 May 2009

Keywords:Motor taskJumpTaking-off and landing modesDevelopment

1050-6411/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.jelekin.2009.05.005

* Corresponding author. Tel.: +33 1 56 56 12 22; faE-mail address: [email protected] (E. Golome

a b s t r a c t

Imitative behavior underlaid by perception and action links during children’s development in complexlocomotor skills has been the object of relatively few studies. In order to explore children’s motor coor-dination modes, 130 children divided into five age groups from 3.5 to 7.5 years were instructed to imitatejumping tasks in spontaneous motor situation and in various imitative contexts by an adult providingverbal orders and gestural demonstrations. Their conformity to the model, stability and variability scoreswere coded from a video analysis when they performed jumps with obstacles. To evaluate their postural-motor control level, the durations of the preparatory phase and jumping flights were also timed. Resultsshowed that all age groups generated the demonstrator’s goal but not necessarily the same coordinationmodes of jumping. In imitation with temporal proximity, the model helped the youngest age groups toadopt his coordination modes and stabilized only the oldest age groups’ performances starting from5.5 years old, without effect on learning imitation. Differences between the youngest and oldest childrenin the jump duration suggested that the reproduction of a complex motor activity such as jumping with aone foot take-off would require resolution and adjustment of main postural stability.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Interference between action observation and action executionwas demonstrated in numerous behavioral studies (Meltzoff andDecety, 2003; Kilner et al., 2003; Prinz, 2005). The basic process in-volved several paradigms allowing the reproduction of hand orfoot movements performed by another person to achieve a goal(Fadiga et al., 1995; Iacoboni et al., 1999; Rizzolatti et al., 2001;Rumiati and Bekkering, 2003; Buccino et al., 2004). A few studiesinvestigated imitative complex motor skills among children(Cadopi et al., 1996) but fewer in jumping activities. Due toembodiment differences between adult models and children, whendealing with environmental constraints and motor skill levels(Newell, 1986), the mapping between observation and productioncould not be direct (Jackson and Decety, 2004). It was shown thatchildren acquired a wide variety of novel actions via imitative mo-tor learning (Bandura, 1977; Abravanel and Gingold, 1985). Butvery often a mere copy of the behavior surface displayed by anadult might not be appropriate or might even be impossible forchildren to perform, due to different limb lengths, body sizes, per-spectives and available skills in motor repertory (Prinz, 2005; Bux-baum et al., 2000; Erlhagen et al., 2006). Before the age of eight,children feel more difficulties in representing their body segments.This a priori makes it difficult for them to imitate any model dem-onstrations (Deloache et al., 2004). Thus, the present study ex-

ll rights reserved.

x: +33 1 56 56 12 12.r).

plored children’s capacity to imitate a complex motor task atdifferent age levels.

As to the main constraints, Assaiante et al. (2005) suggestedthat the first step for children to control their locomotor activitieswas to build a repertory of postural strategies and then, to selectthe most appropriate postural mode, depending on their abilityto maintain equilibrium and to exert force. In jumping, the twoessential functions involved a compromise between body propul-sion and maintenance of whole-body stability (Assaiante et al.,1997). Other jumping constraints were specifically linked totake-off or landing modes (one foot versus two feet) and to theground environment (flat ground versus ground with obstacles)or to the previous actions associated with the jump (Vaivre-Douretand Bloch, 1995; Labiadh et al., 2003). To solve these constraints,children attempted to manage postural adjustments according totheir developmental calendar. Thus, from 21 to 36 months, themain difficulties of spontaneous jumps were to acquire the thrustnecessary for the support foot or feet to become airborne and toorganize the stopping movement upon landing (Vaivre-Douretand Bloch, 1995). For 6 and 7 year old children, the transition inthe organization of balance control took place at the level of lowerlimb coordination during the take-off preparatory phase (Assaianteet al., 1997). Another parameter involved in jumping consisted ofvertical forces exerted on the lower limbs (Jensen and Phillips,1991).

Concerning the ontogenesis of jumping, Jensen et al. (1994)demonstrated the appearance of a mature pattern of jumping coor-dination at the earliest stages of behavior (3–4 years old), requiring

http://dx.doi.org/10.1016/j.jelekin.2009.05.005

mailto:[email protected]

http://www.sciencedirect.com/science/journal/10506411

http://www.elsevier.com/locate/jelekin

L. Labiadh et al. / Journal of Electromyography and Kinesiology 20 (2010) 322–329 323

however a gradual evolution of the process. The emergence ofjumping was linked to solving two main constraints – equilibriumand propulsion. From a developmental point of view, Paoletti(1999) distinguished three conventional jumping types, with eachtype divided into three stages. (1) The drop jump enabled the chil-dren aged under 5 years old to go down the last stair of a staircase.Initially, there was no clear jump but rather a long stride. Then, thetake-off took place simultaneously on both feet. Equilibrium whenlanding on both feet was unstable. (2) The long jump concernedchildren aged 3–5.5 years. At first the children used their arms tohelp them to maintain balance during flight. The landing was stiffand the children were off-balance. Finally, they gave a completepush forward with their legs while swinging their arms. (3) Lastly,the vertical jump concerned children aged 4–6 years. At first, take-off on both feet was badly synchronized and landing was unsuit-able. During flight, full body extension was not complete. Finally,the children performed a synchronized flight with leg greater flex-ion to prepare take-off.

From this literature investigating constraints and gradualdevelopment in jumping, the fact that a child is precociously ableto imitate a model gesture does not mean that this child will beable to reproduce the kinematics details of the movement or thewhole movement. A child will copy the action but not the detailsof the movement. The imitation is not concerned with an accuratekinematic gesture, but rather with a global morphological organi-zation, an action (e.g. taking a glass, lifting an arm) (Deloacheet al., 2004).

From this literature, the first hypothesis was that imitativematching between adult model demonstration and child perfor-mance would be defined in terms of goals – succeed in the jumpregardless of coordination means or modes. Knowing that anadult’s biomechanical and motor jumping repertory was differentfrom a child’s, the second hypothesis was that adult model coordi-nation would not influence child behavior whatever the imitativecontext. However, it could have an influence when children reacha more advanced jump control repertory.

To check these assumptions, a developmental approach ex-plored children’s imitative performances from 3.5 to 7.5 year oldsand their motor control capacity to imitate jumping tasks froman adult experimenter’s verbal orders and gestural demonstra-tions. Their imitative performances were coded from a video anal-ysis when children performed vertical jumps, then drop jumpswith obstacles.

2. Method

The experiment was realized by the adult experimenter in aspontaneous motor situation and in imitative contexts. The sponta-neous motor situation allowed various levels of children’s motorcontrol to be revealed and a referential basis compared to the imi-tative contexts to be established. The locomotor activity was cho-sen because it could be precociously monitored (Paoletti, 1999).The jumping task was adopted from the beginning and performedwhatever the imitative circumstances.

This study had been approved by the Ethics Committee of theParis Descartes University.

Forward N°2 n°3

n°1 N°1 25 cm

Vertical jump Drop jum

Walking On obstacles

Wal

Fig. 1. Experimental course of tasks in a spontan

2.1. Participants

The experimental group was instructed by the adult model toimitate jumping only in the imitative contexts. It was composedof 85 children divided into five age groups (3.5, 4.5, 5.5, 6.5 and7.5 year olds). Each age group was comprised of 17 children. Theexperimental group performed the task from the gestural demon-stration of the adult experimenter. Another group marked as thecontrol group was formed in order to avoid possible interactionswith the experimental group because all the children belonged tothe same school and were likely to meet each other during the en-tire duration of the experiments. Thus, there would be no commu-nication between the experimental and the control groups topreserve the behavioral spontaneity of the control group. The con-trol group was composed of 45 children also divided into five agegroups – 3.5–7.5 years. Each group was comprised of nine children.These children reproduced the same tasks as the experimentalgroup only in the spontaneous motor situation from a verbal ordergiven by the same adult experimenter. Each child from each agegroup performed the task alone with the adult experimenter.Authorizations were obtained from the children’s parents, theschool director and the teachers.

2.2. Material

The children’s performances were filmed both forward to back-ward with one video camera (JVC, 25 images/s) fixed to the groundand operated by a cameraman who followed their reproductions inthe sagittal direction (see Fig. 1).

Three 30 cm diameter bowls were placed in the experimentalcourse in various places. Two were put at each extremity of thecourse (n�1 and n�3) and were used to materialize the departureand the arrival of the locomotor activities. Bowl n�2 was put inthe middle of the course and used as a landing strip after jumping.

Four 30 cm long, 20 cm wide and 10 cm high obstacles were putin pairs between bowls n�1 and n�2, n�2 and n�3. Each obstacle wasseparated by 25 cm used for the reproduction of the jumpingmodes: vertical jump, then drop jump.

2.3. Protocol

The jumping task was preceded by a walk because the original-ity of this research was to produce linking motor actions in acourse. This course took place in a school sports room. There wereneither spectators nor other activities during this experiment. Apreliminary experiment was carried out by each age group tocheck the compatibility of each set of tasks (vertical then dropjumps). The model order in the imitative contexts was: «look atme and then try to do exactly the same thing».

At forward, the model started with his feet in bowl N�1 andwalked on the two first obstacles. Arriving at the second obstacle,he jumped by taking-off on one foot and landing on both feet inbowl N�2. From this bowl, he walked forward placing each foot be-tween the two last obstacles. Arriving in front of the last obstacle,he jumped across by taking-off on one foot and landing on bothfeet in bowl N�3.

n°2

N°3 N°4

Backward

p

king Between obstacles

eous motor situation and imitative contexts.

Table 1Children’s successful scores (%) according to the age group (3.5, 4.5, 5.5, 6.5 and7.5 year olds) in one foot take-off (left part) and both feet landing jump (right part) ina spontaneous motor situation (SMS).

One foot take-off (%) Two feet landing (%)

3.5 years 0 904.5 years 45 905.5 years 55 1006.5 years 45 777.5 years 100 66

324 L. Labiadh et al. / Journal of Electromyography and Kinesiology 20 (2010) 322–329

At backward, the children were to reproduce the same pre-scribed tasks as forward.

To establish a spontaneous motor situation (SMS), only the con-trol group performed the jumping task from a verbal order givenby an adult experimenter for one trial.

The experimental group performed in the five imitativecontexts:

(1) Immediate imitation in the same imitative course (IISIC): theadult model and the child performed immediately in thesame imitative course for one trial.

(2) Simultaneous imitation in two parallel imitative courses(SI//IC): the model and the child performed at the same timebut each one in his own course for one trial.

(3) Time-lag imitation (T-LI): just after finishing their simulta-neous imitation, the children reproduced the jumping inthe same imitative course without the model accompanyingfor one trial.

(4) Deferred imitation (DI): the children imitated this task aftera model demonstration trial without the model accompany-ing them for one trial where the delay separating the modeldemonstration and the children reproduction were from fiveto fifteen seconds.

(5) Learning imitation (LI): the adult model demonstrated thetask at the beginning of each of six sessions, just after theadult model finishing, the children reproduced the jumpingin the same imitative course without the model accompany-ing. Each session was separated by a one week interval. Thechildren reproduced the same task in random order thus, ina different order each week.

2.4. Data collection and statistical analysis

2.4.1. The realization form parametersFor all the children placed in the spontaneous motor situation

as well as in the imitative contexts, the realization forms (one footor two feet) were coded in binary data (1–0) by the adult modelhim-self, then transformed into percentages (%). The performancepercentages were analysed according to overall imitative scoresand to three deliberateness indicators of conformity to the model:(a) the stability to conformity (StCf) – score P 80%: the childrenimitated the model exactly; (b) the stability to nonconformity(StNCf) – score 6 20%: the children imitated the model differently;(c) the variability (Varia) – score between 20% and 80%: the chil-dren oscillated between several coordination modes. These perfor-mances were related to the global number of children for each agegroup. The statistical process of binary data mobilized specificmethods. The percentages were not shared out according to a nor-mality law, and thus did not allow the normality test. So, it wasnecessary to use a transformation of variables to apply an adequateANOVA/MANOVA. The most frequently used transformation –angular transformation of percentages (Lellouch, 1996) was per-formed. The software used for data analysis was Statistica 6.1�

(Statsoft, Inc.).A pairwise post hoc analysis was carried out to detect significant

interactions between the experimental variables. So, the ReducedDistance test compared the imitative context and age group foreach context. The statistical significance level was set at p < 0.05.

The three dependent variables were the children’s successfulperformances in the realization forms: conformity scores (%); sta-bility scores (%) and individual variability scores (%).

The three independent variables were: the five age groups –3.5–7.5 years, the two jumping types – vertical and drop jumpsand the two directions of task production – forward andbackward.

2.4.2. The postural-motor parametersIn order to evaluate the whole-body postural-motor levels of

control for each age section, the jumps were timed for two jumpingphases: the preparatory and the flight phases. The preparatoryphase started at the end of the walking step and finished whenthe foot or feet left the ground for landing. The flight phase corre-sponded to the time elapsed between the moment when the foot/feet left the ground or obstacle and the moment when they landed.This quantitative parameter – the duration measured in seconds,considered as the dependent variable, was submitted to a varianceanalysis and then to an adequate post hoc analysis by pairwisecomparisons. The test of planned comparisons was carried out todetect significant interactions between the experimental variableswith a level set at p = .05 throughout the statistical tests.

2.4.3. The plan of the variance analysesFor each motor situation (the spontaneous and various imita-

tive contexts) an ANOVA variance analysis was realized.For the spontaneous motor situation, this analysis was a two-

factor ANOVA, composed of the independent between-participantfactors – the age group (five levels: 3.5–7.5 years) and the twojumping types (two levels: vertical and drop jumps).

For the four imitative contexts, this analysis was a four-factorANOVA, the independent between-participant factors were theage group (five levels: 3.5–7.5 years), the imitative contexts (fourlevels: IISIC, SI//IC, T-LI, DI) and the two jumping types (two levels:vertical and drop jumps) and the repeated within-participant fac-tor was the execution way (two levels: forward and backward).

For the learning imitation over the six sessions (LI), this analysiswas a three-factor ANOVA, the independent factors were the agegroup (five levels: 3.5–7.5 years) and the two repeated within-participant factors were the trial factor (six levels of repetitions)and the execution way factor (two levels: forward and backward).

3. Results

3.1. Spontaneous motor situation (SMS)

For the one or two footed take-off mode, variance analysisshowed a significant effect of the age factor: F (4,+inf) = 11.20;p < 0.0001. The post hoc analyses showed differences from the3.5 year olds and the other age groups (cf. Table 1 and (Fig. 2)).The same analysis showed that the 7.5 year olds were significantlydifferent from the 4.5, 5.5 and 6.5 year olds on taking-off. However,the children’s landing modes were comparable at all ages. Therewas no significant effect between the two jumping types (dropand vertical jumps) (p > 0.05). The 3.5 year olds systematicallyused both feet while taking-off as well as landing (100%). This tech-nique was shared by the 4.5 and 5.5 year olds only while landing.All the 7.5 year olds used one foot to take-off (100%).

3.2. Imitative contexts in the IISIC, SI//IC, T-LI, DI

For the overall scores, variance analysis showed a significant ef-fect of the age factor: F (4,+inf) = 17.40; p < 0.0001. The Reduced

3.5 yeras

4.5 years

5.5 years

6.5 years

7.5 years

Suc

cess

ful s

core

s (%

) in

"S

MS

"

0.0

0.2

0.4

0.6

0.8

1.0

teefhtobgnidnaLtoofenoffo-gnikaT

Fig. 2. Variance analysis results of success scores (%) showing significant differenceaccording to the age group (3.5, 4.5, 5.5, 6.5 and 7.5 year olds) in take-off (left part)and no significant difference in landing jump (right part) in a spontaneous motorsituation (SMS).

3.5 years

4.5 years

5.5 years

6.5 years

7.5 years

Children scores in imitative contexts (ICTP)

Succ

essf

ul s

core

s (%

) in

ICTP

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

IISIC SI//IC T-LI DI

Fig. 3. Variance analysis results of success scores (%) for jumping showingsignificant difference between the imitative contexts: immediate imitation in thesame imitative course (IISIC), simultaneous imitation in two parallel imitativecourses (SI//IC), time-lag imitation (T-LI), deferred imitation (DI) and the children’sgroups (3.5, 4.5, 5.5, 6.5 and 7.5 year olds).


Distance test revealed that the 3.5 and 4.5 year olds performed low-er conformity scores than the other age groups. There was no sig-nificant interaction between the age and the imitative contexts: F(12,+inf) = 0.74; p > 0.05.

Other variance analysis concerning each age section showed asignificant difference between the four imitative contexts onlyamong the two younger age groups (3.5 and 4.5 year olds). The posthoc analysis was significant (p < 0.05). The first three imitative con-texts had better scores than the deferred imitation (DI) context (cf.Table 2 and Fig. 3)). The results of the older children (5.5 to 7.5 yearolds) did not show different scores between imitative contexts.Whatever the age group, for the children’s overall scores, verticaljumps (score 1) revealed no difference from drop jumps (score 2)(cf. Table 2).

Concerning the stability of conforming to the model, almost halfof the 3.5 years old (47%) were stable to their own coordinationmodes stability to the nonconformity (StNcf) and the second half(48%) was variable (Varia) in adopting take-off and landing modes.Only 5% of these children conformed to the model coordinationmode (StCf). As for the 4.5 year olds, 52% systematically adoptedthe same mode as the 3.5 year olds: stability to the nonconformity(StNcf) (two feet/two feet or one foot/one foot). Lastly, forty-threepercent of these children used two modes (Varia) and only 5% were

Table 2Children’s overall scores (%) in each imitative context: immediate imitation in the same imtime-lag imitation (T-LI) and deferred imitation (DI). The average scores (Av.Sc) of IISICconformity to the model (right part): stability to conformity (StCf), – variability (Varia) andscore 2 (S2) = coordination mode – drop jump.

Overall scores

Scores IISIC (%) SI//IC (%) T-LI (%)

3.5 years S1 29 35 39S2 38 24 47

4.5 years S1 26 41 26S2 24 21 38

5.5 years S1 68 76 54S2 59 53 57

6.5 years S1 56 74 65S2 59 47 37

7.5 years S1 71 76 59S2 56 79 59

stable to conformity to the model. For the 5.5 year olds, 54% werevariable (Varia) but 29% started to be stable (StCf) using one foot intaking-off and both feet in landing. The rest of these children (17%)were stable to their own conformity (StNcf). The 6.5 year olds chil-dren showed the most variability (84%). The 7.5 year olds childrenaccomplished higher stability (StCf) to conformity to the model(41%) than the other age groups (cf. Table 2).

3.2.1. Link between walking and jumping in IISIC, SI//IC, T-LI, DIConcerning the link between walking and jumping, the chil-

dren’s overall scores indicated from 3.5 to 7.5 years, 18%, 44%,68%, 76%, 67% averaged for the four imitative contexts (IISIC, SI//IC, T-LI, DI). Similarly, stability to conformity to the model (StCf)improved progressively with each age – 0%, 11%, 58%, 47%, 64%.

3.2.2. Preparatory phase duration in IISIC, SI//IC, T-LIVariance analysis showed a significant global effect of the age fac-

tor – F (1,4) = 5.9, p < 0.0001 and of the imitative context factor – F(1,4) = 6.3, p < 0.01. However these factors were linked becausethere was a triple significant interaction between the jumping type,the imitative context and the age factors – F (1,79) = 1.9, p < 0.05.Post hoc analyses by pairwise comparisons (planed comparison test)showed that the 3.5 year olds were significantly slower than the

itative course (IISIC), simultaneous imitation in two parallel imitative courses (SI//IC),, SI//IC, T-LI (left part) for each age group (3.5, 4.5, 5.5, 6.5 and 7.5 year olds). The– stability to nonconformity (StNCf). Score 1 (S1) = coordination mode – vertical jump,

Conformity to the model

DI (%) Av.Sc (%) StCf (%) Varia (%) StNcf (%)

15 34 5 48 4718 3615 31 5 43 5221 2762 66 29 54 1753 5665 65 11 84 538 4838 69 41 47 1159 65

Table 3Children’s overall scores in the duration of the preparatory phase of the jump for thethree imitative contexts: immediate imitation in the same imitative course (IISIC),simultaneous imitation in two parallel imitative courses (SI//IC) and time-lagimitation (T-LI) and for the five age groups (3.5, 4.5, 5.5, 6.5 and 7.5 year olds).

Overall scores in each imitative contexts (s)

3.5 years: IISIC = 1.61 ± 1.38 SI//IC = 1.28 ± 1.52 T-LI = 0.97 ± 0.664.5 years: IISIC = 0.90 ± 1.11 SI//IC = 1.06 ± 1.41 T-LI = 1.09 ± 1.145.5 years: IISIC = 0.62 ± 0.47 SI//IC = 0.64 ± 0.62 T-LI = 0.73 ± 0.746.5 years: IISIC = 0.77 ± 0.2 SI//IC = 0.57 ± 0.08 T-LI = 0.67 ± 0.087.5 years: IISIC = 0.73 ± 0.44 SI//IC = 0.48 ± 0.07 T-LI = 0.64 ± 0.23

Table 4Children’s overall scores in the flight duration of the jump for the three imitativecontexts: immediate imitation in the same imitative course (IISIC), simultaneousimitation in two parallel imitative courses (SI//IC) and time-lag imitation (T-LI) andfor the five age groups (3.5, 4.5, 5.5, 6.5 and 7.5 year olds).

Overall scores in each imitative contexts (s)

3.5 years: IISIC = 0.32 ± 0.18 SI//IC = 0.31 ± 0.12 T-LI = 0.31s ± 0.144.5 years: IISIC = 0.34 ± 0.24 SI//IC = 0.30 ± 0.17 T-LI = 0.29s ± 0.615.5 years: IISIC = 0.33 ± 0.11 SI//IC = 0.34 ± 0.32 T-LI = 0.28s ± 0.116.5 years: IISIC = 0.29 ± 0.05 SI//IC = 0.30 ± 0.17 T-LI = 0.18s ± 0.197.5 years: IISIC = 0.24 ± 0.02 SI//IC = 0.30 ± 0.25 T-LI = 0.28s ± 0.14

Table 5Children’s overall scores (%) for jumping in deferred imitation (DI) (left part) for eachage group (3.5, 4.5, 5.5, 6.5 and 7.5 year olds). The conformity of the model (rightpart): stability to conformity (StCf), – variability (Varia) and – stability to noncon-formity (StNCf). Score 1 (S1) = coordination mode – vertical jump, score 2 (S2) = coor-dination mode – drop jump.

Overall scores Conformity to the model

StCf (%) Varia (%) StNcf (%)

3.5 years S1 12% 0 18 82S2 11%

4.5 years S1 35% 0 71 29S2 26%

5.5 years S1 66% 11 89 0S2 54%

6.5 years S1 60% 5 95 0S2 52%

7.5 years S1 70% 17 83 0S2 61%

3,5 years

4,5 years

5,5 years

6,5 years

7,5 years

Taking off and landing children sections/six sessions

Succ

essf

ul s

core

s (%

) in

lear

ning

imita

tion

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

S1 S2 S3 S4 S5 S6

Fig. 4. Variance analysis results of success scores (%) for jumping in learningimitation (LI) showing significant differences between six sessions (S1–S6) aver-aged at forward/backward for each age group and the children’s groups (3.5, 4.5,5.5, 6.5 and 7.5 year olds).


other age groups, but only for the vertical jump and in the two imi-tative contexts than in the T-LI: for IISIC – F (1,79) = 7.5, p < 0.01and for the SI//IC – F (1,79) = 24.2, p < .0001. (cf. Table 4).

3.2.3. Flight phase duration in IISIC, SI//IC, T-LIVariance analysis showed a global significant effect of the age

factor – F (1,4) = 854.9, p < 0.0001: the 3.5 year olds were signifi-cantly slower than the other age groups and of the jumping typefactor – F (1,4) = 6.04, p < 0.01: the vertical jumps were slower thanthe drop jumps (Table 5). Furthermore, there was a significant dou-ble interaction between the imitative context factor and the execu-tion way factor – F (2,8) = 5.4, p < 0.01. The post hoc analysesshowed that all the children were slower for the forward thanfor the backward execution way only for the immediate context– F (1,80) = 11.4, p < 0.001. In the backward execution way, theduration was slower for the immediate context than for thetime-lag context – F (1,4) = 7.14, p < 0.01 and for the spontaneouscontext – F (1,4) = 5.07, p < 0.02 but not between the simultaneousand the time-lag imitation contexts – F (1,80) = 2.27, p < 0.13.

3.3. Learning imitation (LI) in six sessions

Variance analysis showed a significant global effect of the agefactor – F (4,+inf) = 53.78; p < 0.0001. The Reduced Distance test re-

vealed the difference of the youngest children (3.5 and 4.5 years)who obtained lower scores than the other age groups. There wasno significant interaction between ages and sessions: F(20,+inf) = 1.29; p > 0.05 (cf. Table 3 and Fig. 4).

Concerning conformity to the model, the 3.5 year olds systemat-ically used two jumping modes (82%) (two feet/two feet mode (54%)and one foot/one foot mode (28%)). The other children groups oscil-lated between some modes – 71%, 89%, 95%, 83%, progressivelyreaching the two feet/two feet mode – 38%, 19%, 8%, and 7%.

3.3.1. Link between walking and jumping in LIConcerning the link between walking and jumping, the chil-

dren’s overall scores showed respectively age per age from 3.5 to7.5 year olds – 17%, 44%, 76%, 83%, and 87%, averaged for the sixsessions at forward/backward directions. Similarly, the stabilityto conformity to the model (StCf) improved progressively withage (0%, 0%, 35%, 64%, and 88%).

3.3.2. Preparatory phase duration of the jumps in LIVariance analysis showed a significant effect of the age factor –

F (1,4) = 17.62, p < 0.0001, of the jumping type – F (1,4) = 9.75,p < 0.01, of the execution way – F (1,4) = 29.42, p < 0.0001 and ofthe session – F (1,4), p < 0.0001. However these factors were linkedbecause there was a significant triple interaction between the exe-cution way, the age and the session factors – F (20,395) = 1.62,p < 0.01. The post hoc analyses by pairwise comparisons showedthat the youngest children jumped slower than the other childrenfor the forward execution way and the three first sessions. Thesame analysis showed a significant triple interaction betweenjumping type, execution way and age factors – F (4,79) = 2.72,p < 0.01. The post hoc analyses showed a difference to the 3.5 and4.5 year olds who spent more time than the other age groups forthe vertical jump and for the forward direction. The durationscores of the 5.5, 6.5 and 7.5 year olds were similar. There was asignificant triple interaction between jumping type, executionway and session factors – F (5,395) = 2.73, p < 0.01. The post hocanalyses revealed that all the children decreased this time in thecourse of the sessions for the forward execution and the verticaljumps (cf. Table 6).

Table 6Children’s overall scores in the duration of the preparatory phase of the jump for learning imitation (LI) in the six sessions averaged on forward/backward for each age group.

3.5 years 4.5 years 5.5 years 6.5 years 7.5 years

Six sessionsonforward

Six sessionson backward

Six sessionson forward








1.13 ± 0.72 s 1.07 ± 0.75 s 1.01 ± 0.48 s 0.80 ± 0.63 s 0.70 ± 0.40 s 0.60 ± 0.45 s 0.68 ± 0.10 s 0.56 ± 0.13 s 0.54 ± 0.03 s 0.54 ± 0.09 s

Table 7Children’s overall scores in the flight duration for learning imitation (LI) in the six sessions averaged on forward/backward for each age group.

3.5 years 4.5 years 5.5 years 6.5 years 7.5 years

Six sessions averaged onforward/backward





0.3 s ± 0.1 0.33 s ± 0.26 0.27 s ± 0.1 0.28 s ± 0.07 0.28 s ± 0.14


3.3.3. Flight phase duration in LIVariance analysis showed a global significant effect of the age

factor in flight duration – F (1,4) = 3.13, p < 0.01 and of the execu-tion ways – F (1,4) = 6.2, p < 0.01. The 3.5 and 4.5 year olds spentmore time than the other age groups and all children spent moretime for the forward execution way than for the backward way.The same variance analysis showed a significant triple interactionbetween the jumping type, the session and the execution way fac-tors – F (5,395) = 2.24, p < 0.05. The post hoc analyses revealed thatall the children spent more time in the first three sessions, the ver-tical jump and the forward execution way (cf. Table 7).

4. Discussion

First hypothesis was that the imitative matching between childand adult model would be defined in terms of goal. Results verifiedthis assumption. All the children walked then jumped without anyother motor behavioral goal such as reaching obstacles or carryingobjects (Byrne, 1999; Wohlschläger et al., 2003). This result wasconsistent with Johansson’s work (1973) demonstrating humancapability to recognize biological motions from a small numberof structured visual cues. But it is not known if this knowledge isestablished only by simulation on the basis of similar motor cir-cuitry (Rizzolatti et al., 2001) or by semantic categorisation (Jeann-erod, 1999). These associations could be related to the functions ofmirror neurons linked to the reproduction and the observation ofhand targets and pedestrian skills (Buccino et al., 2004). In a previ-ous study, Perrett et al. (1990) demonstrated that the superiortemporal sillon (STS) was activated when observing acts involvingfull body movements such as locomotor activities (walking orjumping). Since STS is activated by the display of learned motorskills such as a golf swing or upside-down walking (Blakemoreand Decety, 2001), and feet mobilization (Buccino et al., 2004), thisregion could participate in jumping knowledge. So while observingother persons with the intention of imitating their actions, the fulldetails of their motions are not likely to be encoded but those mo-tions could often be interpreted in terms of a demonstrator’s goal(Erlhagen et al., 2006).

However, the integration of the coordination modes could de-pend on the children’s degree of control over their own motor rep-ertory. This assumption was verified by our results. (a) Theyoungest age groups (3.5 and 4.5 years old) showed some difficul-ties in solving the jumping constraints with one foot while taking-off. Observable findings showed that the two feet/two feet strategywas used more in spontaneous motor situations and imitative con-texts (Vaivre-Douret and Bloch, 1995). The propulsive force andequilibrium constraints could prevent children from adopting theone foot approach in taking-off (Vaivre-Douret and Bloch, 1995).In walking tasks, children used both feet to overstep each obstacle

(Labiadh et al., 2003). This behavior could justify their two footedtake-off mode. They were evenly resigned to a difficulty in thetwo linked locomotor activities (walking and jumping). The 4.5and 5.5 year olds randomly used one foot or both feet while tak-ing-off and both feet while landing. The qualitative performancesof the 6.5 year olds showed that they had more flexibility in orga-nizing their jumping modes. They chose one foot or both feet whiletaking-off and both feet upon landing. The 7.5 year olds had no dif-ficulty using one foot for taking-off (100%). However their landingmodes varied between both two feet and one foot. These resultsare coherent with the literature data that states that in a spontane-ous motor situation, no child of 3.5 years (9 children here) used ajumping mode other than with two feet/two feet. In our study, arate of about 30% was reach from the first demonstration that sta-bilized in (ICTP), then decreased by 11–12% in learning imitation.Authors showed the importance of three motor control factors:the feet pushing to take-off; the postural organization to dropwhen landing and the intersegmental coordination of the lowerlimbs for all the phases (Vaivre-Douret and Bloch, 1995; Assaianteet al., 2005). This may mean that the sight of a synchronizationmode might not influence what children are really able to do.

The second hypothesis assumed that the adult model’s coordi-nation would not influence the children’s behavior whatever theimitative context but, that it could influence them when theyhad acquired a more advanced jump control repertory. Indeed,the results showed the following (1) in IISIC and SI//IC, the youn-gest children were helped by the model for more conformity intake-off, compared to differed imitation when the model did notguide the children’s imitative performances. The gestural model‘‘forced” the children to adopt his expression regardless of theirdevelopment level. Even if the children had the ability to expressthis gesture externally, they did not perform it spontaneously. In-deed, the two footed take-off mode was not easily adopted bythe youngest children, considering the constraints of asymmetrictake-off compared to the use of both feet to obtain propulsionand equilibrium (Jensen et al., 1994). (2) The model had an inter-ference effect between 5.5 and 7.5 year olds only in stability termswhere it was perceptible in ICTP respectively (29% and 41%) com-pared to their stability in learning imitation (LI) respectively (11%and 17%). The 4.5 and 6.5 year olds reproduced this jumping withcomparable scores in the imitative contexts. (3) Moreover, theenvironmental constraints of the materials could be a pregnantfactor intervening in the production of the link between walkingand jumping. This assumption was intensified because of someindications related to the gestural form: (i) a fall in individual sta-bility at 7.5 years in LI compared to ICTP, (ii) a fall of stability occu-red in the same context for 5.5 years; (iii) a fall in performancestook place between ICTP and LI for 3.5 years; (iv) there was stabil-ity of the same sign between 4.5 and 6.5 years regardless of the


imitative contexts; (v) no learning effect appeared in any age groupand for all the six sessions. The quantitative results concerning thelink between walking and running was possible because the modeldemonstration informed and induced this link from the first obser-vation so that the children were able to follow the model’s strat-egy. Similarly, in the two jumping modes (drop and verticaljumps), the child would necessarily need his full body to drop,while in a vertical jump, slight landing or a simple coming downwithout strong extension, would be sufficient to reach the bowl.Thus a ‘‘real” jump could be attested for the youngest children(3.5 and 4.5 year olds) and this would be consistent with the factthat the 6.5 and 7.5 and to a lesser degree the 5.5 year olds usedone foot take-off and two feet or on foot landing.

The time data of the flight and preparatory jumping phases con-firmed that the children’s imitative behaviors varied with the pos-tural-motor development as the second part of the hypothesisenvisaged it. Results concerning these two durations showed nodifference between imitative contexts only for the older children.This was coherent with their scores in demonstrated gesturalreproduction. It revealed an immediate and a quasi systematicadoption of the one foot landing strategy accompanied by fluidmotor coordination. This older children’s behavior would meanthat the moving modes of the model were immediately understoodthen, influenced however by the view of the obstacles. These re-sults are corroborated by the fact that direct observation of theproduction model was facilitated (Hommel et al., 2001; Prinz andHommel, 2002). Recent literature data confirms all these findings.The 3.5 year olds were sensitive to the reading of a linking meansthat mobilized their body parts. But to succeed in motor coordina-tion, the children might solve the main constraints and building di-rect and indirect models to predict and control their motorcommands (Schaal et al., 2003). The older children (7.5 year olds)seemed to control the required coordination, almost solving thepostural adjustments. This result was compatible with the Assaian-te et al. (2005) findings suggesting that children aged from 6 to7 year olds reached a crossroads in both postural and equilibriumlocomotor control.

Results concerning the postural-motor parameters with severalfactor interactions confirmed that the children aged under 5 yearsold spent more time in the preparation and the flight duration ofthe vertical jump, for the forward execution way and in the imme-diate imitative context than for the drop jump, the backward exe-cution way and in the other types of imitative contexts.Furthermore, when the learning imitation has been carried out,these children were also slower for the three first sessions thanfor the last sessions.

The preparatory and the flight phase durations were slower forthe vertical than for the drop jumps. As described previously, thejumping type and its execution difficulty (cost of balance and pro-pulsion) is influenced by the development phases: the drop jumpconcerned the children aged under 5 years old and the verticaljump concerned children aged 4–6 years (Paoletti, 1999). The dura-tion of drop and vertical jumps varied according to the imitativecontext: the youngest children spent time than the other groupswith the temporal proximity of the model than without the modelaccompanying. This was compatible with the coordination modesquality previously evoked but also linked with the learning effectwhere the imitation process is one of the first phases. The idea ofthe involvement of the learning effect in this study is supportedby the whole context of the results. Indeed, inside each session,the forward duration is more timed than the backward one andduring the learning imitation the three first sessions were slowerthan the last one. For response synchronization of motor tasks,children differed from adults early in learning (Savion-Lemieuxet al., 2009), but children as well as adults use more attention forthe first execution than for the others as described for young gym-

nasts in learning of a motor task involving equilibrium (Vuillermeet al., 2001; Vuillerme and Nougier, 2004). Moreover, the youngestchildren difficulties seemed to be related to problems in dividingtheir attention between self-focus and perception of the modelperforming. Typically, an individual switches his/her attentionfrom the self to the other person numerous times during the imi-tation (Meltzoff, 2005; Marton, 2009). Concerning the childrenaged under 5 years old, their maturation level of this cognitivequality is not sufficient to allow the task execution.

It has previously been argued that imitating an action could in-volve inversing the model to map the consequences of action con-straints (Wolpert et al., 2003). The output to control motorcommands would be required to achieve or to maintain goals (Jell-ema and Perrett, 2006). The findings of this study revealed a slowevolution of the jumping coordination modes, more importantthan the one noticed in walking (Labiadh et al., 2003).

In sum, the aim of our study was to give an overview of how amodel’s actions such as jumping are parsed, represented andreproduced by preschool children. The demonstrator’s behaviorcould influence the children to imitate goals but not necessarilythe details (modes) of these goals. The differences between theyoungest and oldest children in jump duration suggested that theimitation of a complex motor activity such as jumping with a onefoot take-off would require the resolution of main postural stabil-ity and adjustments.

Acknowledgements

The authors thank the children and their teachers who took partin this study, Michel Bonneau (the cameraman) for his help duringthe experimentations, Sylvain Hanneton for his help in the statisti-cal analysis. The authors were grateful to Alan Stern for proof read-ing the English of the paper and for his valuable advice.

References

Abravanel E, Gingold H. Learning via observation during the second year of life. DevPsychol 1985;21:614–23.

Assaiante C, McKinley PA, Amblard B. Head–trunk coordination during hops usingone or two feet in children and adults. J Vestib Res 1997;7:145–60.

Assaiante C, Mallau S, Viel S, Jover M, Schmitz C. Development of postural control inhealthy children: a functional approach. Neur Plast 2005;12:263–72.

Bandura A. Social learning theory. Englewood Cliffs, NJ: Prentice Hall; 1977.Blakemore SJ, Decety J. From the perception of action to the understanding of

intention. Nat Rev Neurosci 2001;2:561–7.Buccino G, Vogt S, Ritz A, Fink GR, Zilles K, Frend-Johachim H, et al. Neural circuits

underlying imitation learning of hand action. Neuron 2004;42:323–34.Buxbaum LJ, Giovanetti T, Libon D. The role of dynamic body schema in praxis:

evidence from primary progressive apraxia. Brain Cogn 2000;44:166–91.Byrne WR. Imitation without intentionality. Using string parsing to copy the

organisation of behaviour. Anim Cogn 1999;2:63–72.Cadopi M, Chatillon JF, Badly R. Representation and performance. Reproduction of

from and quality of movement in dance by eight- and 11-year-old novices. Br JPsychol 1996;86(Pt):217–25.

Deloache JS, Uttal DH, Rosengren KS. Sole errors offer evidence for a perception–action dissociation early in life. Science 2004;304:1027–9.

Erlhagen W, Mukovskiy A, Bicho E. A dynamic model for action understanding andgoal-directed imitation. Brain Res 2006;1083:74–188.

Fadiga L, Fogassi L, Pavesi G, Rizzolatti G. Motor facilitation during actionobservation. A mangetic stimulation study. J Neurophysiol 1995;73:2608–11.

Hommel B, Müsseler J, Aschersleben G, Prinz W. The theory of event coding (TEC): aframework for perception and action planning. Behav Brain Sci 2001;24:1–77.

Iacoboni M, Woods RP, Brass M, Bekkering H, Mazziotta JC, Rizzolatti G. Corticalmechanisms of human imitation. Science 1999;286:2526–8.

Jackson PL, Decety J. Motor cognition: a new paradigm to study self-otherinteractions. Curr Opin Neuropsychol 2004;14:1–5.

Jellema T, Perrett DI. Neural representations of perceived bodily actions using acategorical frame of reference. Neuropsychologia 2006;44:1535–46.

Jeannerod M. To act or not to act. Perspectives on the representation of actions.Quart J Exp Psychol 1999;52A:1–29.

Jensen JL, Phillips SJ. Variations on the vertical jump: individual adaptation tochanging task demands. J Motor Behav 1991;23:63–74.

Jensen JL, Phillips SJ, Clark JE. For young jumpers, differences are in the movement’scontrol, not its coordination. Res Q Exerc Sport 1994;65:258–68.


Johansson G. Visual perception of biological motion and a model for its analysis.Percept Psychophys 1973;14:201–11.

Kilner JM, Paulignan Y, Blakemor SJ. An interference effect of observed biologicalmovement on action. Curr Biol 2003;13:522–5.

Labiadh L, Ramanantsoa MM, Pavis B, Hanneton S. La taille des enfants est-elle unfacteur critique dans la marche avec obstacles? Journées Nationales d’Etude dela Société Française de Psychologie du Sport, Paris, 20–21 mars; 2003.

Lellouch JP. Méthodes statistiques en expérimentation biologiques. Médecine-Sciences. Elammanon; 1996.

Marton K. Imitation of body postures and hand movements in children with specificlanguage impairment. J Exp Child Psychol 2009;102:1–13.

Meltzoff AN, Decety J. What imitation tells us about social cognition: a rapprochementbetween developmental psychology and cognitive neuroscience. Phil Trans R SocLondon B 2003;29:491–500.

Meltzoff AN. Imitation and other minds: the ‘‘like me” hypothesis. In: Hurley S,Chater N, editors. Perspectives on imitation: from neuroscience to socialscience. Cambridge, MA: The MIT Press; 2005. p. 55–77.

Newell KM. Constraints on the development of coordination. In: Wade MG, WhitingHTA, editors. Motor skill acquisition. Dordrecht: Martinus; 1986.

Paoletti R. Education et Motricité: L’Enfant de deux à huit ans. DeBoeck Université;1999.

Perrett DI, Oram MW, Wachsmuth E, Emery NJ. Understanding the visualappearance and consequence of actions. In: Goodale MA, editor. Vision andaction. New Jersey, Norwood: Ablex Publishing Corporation; 1990. p. 163–80.

Prinz W, Hommel B. Common mechanisms in perception and action: attention andperformance XIX. New York: Oxford University Press; 2002.

Prinz W. An ideomotor approach to imitation. In: Hurlley S, Chater N, editors.Perspectives on imitation: from cognitive neuroscience to social science2005;vol. 1. Cambridge: MIT Press; 2005. p. 141–56.

Rizzolatti G, Fogassi L, Gallese V. Neurophysiological mechanisms underlyingunderstanding and imitation of action. Nat Rev Neurosci 2001;2:661–70.

Rumiati RI, Bekkering H. Imitate or not to imitate? How the brain can do it, that isthe question. Brain Cogn 2003;53:479–82.

Savion-Lemieux T, Bailey JA, Penhune VB. Developmental contributions to motorsequence learning. Exp Brain Res 2009;Apr 11 [Epub ahead of print].

Schaal S, Peert A, Billard A. Computational approaches to motor learning byimitation. Phil Trans R Soc London B 2003;29:537–47.

Vaivre-Douret L, Bloch H. The onset of jumping in young children. ANAE1995;35:181–7.

Vuillerme N, Nougier V. Attentional demand for regulating postural sway: the effectof expertise in gymnastics. Brain Res Bull 2004;63:161–5.

Vuillerme N, Danion F, Marin L, Boyadjian A, Prieur JM, Weise I, et al. The effect ofexpertise in gymnastics on postural control. Neurosci Lett 2001;303:83–6.

Wohlschläger A, Gatti M, Bekkering H. Action generation and action perception inimitation: an instance of the ideomotor principle. Phil Trans R Soc London B2003;29:501–15.

Wolpert DM, Doya K, Kawato MA. Unifying computational framework for motorcontrol and social interactions. Phil Trans R Soc London B 2003;29:593–602.

Lazhar Labiadh, PhD in Psychology from Paris Des-cartes University (2004). The aim of his research insport sciences covers imitation, learning and devel-opment of motor skills.

Marie Martine Ramanantsoa is presently Lecturer in
the STAPS (sciences, technical of sport and physicalactivity) department of Paris Descartes University asPhD in psychology.
Eveline Marie Elisabeth Golomer, MD received her
PhD degrees in Neurosciences from Pierre and MarieCurie University (1995) then, her Authorization to bein charge of Research (HDR 2001) from Paris Des-cartes University. Main research includes spectralanalysis of dynamic spontaneous body sways andmotor control of whole body rotations.

Documents

Preschool-aged children’s jumps: Imitation performances