See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/277969913

Development and Validation of a Test Instrument for the Assessment of Basic

Motor Competencies in Primary School

Article in Measurement in Physical Education and Exercise Science · May 2015

DOI: 10.1080/1091367X.2014.998821

CITATIONS READS

125 1,607

3 authors, including:

Christian Herrmann Erin Gerlach

Pädagogische Hochschule Zürich / Zurich University of Teacher Education University of Hamburg
169 PUBLICATIONS 2,272 CITATIONS 102 PUBLICATIONS 1,174 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Christian Herrmann on 09 February 2016.

The user has requested enhancement of the downloaded file.

This article was downloaded by: [188.103.189.14]
On: 19 May 2015, At: 11:54
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Measurement in Physical Education and Exercise

Science
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/hmpe20

Development and Validation of a Test Instrument

for the Assessment of Basic Motor Competencies in
Primary School
a b a
Christian Herrmann , Erin Gerlach & Harald Seelig
a
Department of Sport, Exercise and Health, University of Basel, Basel, Switzerland
b
Faculty of Human Sciences, University of Potsdam, Potsdam, Germany
Published online: 19 May 2015.

Click for updates

To cite this article: Christian Herrmann, Erin Gerlach & Harald Seelig (2015) Development and Validation of a Test Instrument
for the Assessment of Basic Motor Competencies in Primary School, Measurement in Physical Education and Exercise Science,
19:2, 80-90, DOI: 10.1080/1091367X.2014.998821

To link to this article: http://dx.doi.org/10.1080/1091367X.2014.998821

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Measurement in Physical Education and Exercise Science, 19: 80–90, 2015
Copyright © Taylor & Francis Group, LLC
ISSN: 1091-367X print / 1532-7841 online
DOI: 10.1080/1091367X.2014.998821

Development and Validation of a Test Instrument

for the Assessment of Basic Motor Competencies
in Primary School
Christian Herrmann
Department of Sport, Exercise and Health, University of Basel, Basel, Switzerland

Erin Gerlach
Faculty of Human Sciences, University of Potsdam, Potsdam, Germany
Downloaded by [188.103.189.14] at 11:54 19 May 2015

Harald Seelig
Department of Sport, Exercise and Health, University of Basel, Basel, Switzerland

A central aim of Physical Education (PE) is the promotion of basic motor competencies
(“Motorische Basiskompetenzen” [MOBAK]). These are the necessary prerequisites for devel-
oping a physically active lifestyle. Valid test instruments are needed for the evaluation of the effect
of PE. For this purpose, we developed a test instrument for the assessment of basic motor
competencies in first graders. We empirically investigated the construct validity of this
MOBAK test instrument in a study (N = 317, M = 7.0 years, girls: n = 173, boys: n = 144).
The exploratory factor analysis shows a two factorial structure (locomotion, object control), which
is confirmed in a confirmatory factor analysis (CFI = .96, RMSEA = .036, WRMR = .65). In
subsequent analyses, we were able to demonstrate that the calculation of a factor sum value for
each factor is statistically valid, the factorial structure is the same for boys and girls (χ2 [6] = 6.95,
p = .33), and no differential item functioning exists. The MOBAK test instrument is sufficient for
the test-theoretical requirements and is thus suitable for the evaluation of the effect of PE.

Keywords: factor analysis, motor basis competencies, physical education, test development

INTRODUCTION adolescence (Barnett, Morgan, van Beurden, Ball, & Lubans,

2011; Barnett, van Beurden, Morgan, Brooks, & Beard, 2009)
A central aim of physical education (PE) is the development of and early adulthood (Lubans, Morgan, Cliff, Barnett, & Okely,
a physically active lifestyle (Balz, 2009; National Association 2010; Stodden et al., 2008).
for Sport and Physical Education, 2011). Besides attitudes and In German-speaking educational science in sport and exer-
knowledge, motor skills or basic motor competencies also play cise, motor skills are discussed in the context of the avail-
a central role in this process. They constitute essential require- ability of basic motor competencies. These are related to the
ments for taking part in the culture of sport and exercise and concept of fundamental movement skills mainly discussed in
developing a physically active lifestyle (Stodden et al., 2008). English-speaking sport and exercise science. According to
Thus, several studies show that motor skill development in this cultural line of research on competencies, adolescents
childhood correlates with an increased physical activity in must have basic motor competencies in order to take part in a
culture of sport and exercise and to reflect on sports and
exercise as something valuable in life (Gogoll, 2012).
Correspondence should be sent to Christian Herrmann, Department of A majority of studies within school-based PE research define
Sport, Exercise and Health, University of Basel, Birsstrasse 320B, CH-4052 motor performance as the central learning outcome in PE (in the
Basel, Switzerland. E-mail: christian.herrmann@unibas.ch overview McKenzie & Lounsbery, 2013; Metzler, McKenzie,
 ASSESSMENT OF BASIC MOTOR COMPETENCIES 81

van der Mars, Barrett-Williams, & Ellis, 2013; Rink, 2013). techniques. In the related instruments, the MOBAQ test items
This provides a pragmatic perspective for evaluating effects in are constructed on the basis of normative pedagogical discus-
PE. However, no uniform procedure has yet been established, sions which respond to the question of which competencies a
and valid evaluation instruments are required (McKenzie & child should exhibit at a certain age to be able to participate in
Lounsbery, 2013). In PE and exercise research, the basis of PE as well as in sport and exercise culture. The purpose of the
motor performance is discussed from two perspectives: the MOBAQ instruments is to assess learning status (What are
ability-oriented approach (Bös & Mechling, 1983; Gundlach, children able to do at the end of a training program? What
1968; Roth, 1982), which can primarily be found in the special educational needs do children have?). Furthermore,
German-speaking area, and the skill-oriented approach (in the this instrument was developed to control for the fulfilment of
overview Burton & Miller, 1998; Cools, de Martelaer, Samaey, the minimal standards (i.e., curricular outcome).
& Andries, 2009), which is primarily used in the English- Thus, Herrmann, Leyener et al. (submitted) developed
speaking area. Alternatively, the recently developed basic the MOBAQ instrument for seventh grade and tested the
motor qualification (“Motorische Basiskompetenzen” instrument for construct validity. This study revealed a two-
[MOBAQ]) approach (Herrmann, Leyener, Pühse, & Gerlach, factor structure which—similarly to the motor skill assess-
submitted; Kurz, Fritz, & Tscherpel, 2008) can be seen as a third ment (in the sense of fundamental movement skills)—differ-
perspective which is positioned between the two aforemen- entiated between body movements (locomotion) and
tioned approaches. This approach stems from a sport pedago- movements with a ball (object control). Moreover, this
gical tradition. factorial structure proved to be stable over time, and the
Downloaded by [188.103.189.14] at 11:54 19 May 2015

Proponents of the ability-oriented approach assume that test instrument fulfilled common psychometric test criteria.
motor activity is structured into conditional basic character- Although various test instruments have been developed,
istics of strength, endurance, speed, coordination, and flex- none of the existing tests was designed explicitly to evaluate
ibility (Bös & Mechling, 1983). The most prominent test students’ learning performance in PE courses. Such a test
instrument is currently the Deutscher Motorik-Test (German needs to be capable of adequately measuring the motor per-
Motor Test; Bös et al., 2009). Motor abilities are considered formance dispositions required by the curricula (curricular
to be latent, stable (over a period of time) dispositions and validity). In this article, we present the development of an
describe context-free motor proficiency. instrument for assessing basic motor competencies in first
Within the skill-oriented approach (c.f. Burton & Miller, graders (pupils aged 6–7 years) as well as an evaluation of its
1998; Clark & Metcalfe, 2002; Stodden & Holfelder, 2013), construct validity. We are particularly interested in achieving
suitable instruments for the assessment of motor skills are a close link to the curriculum. This will ensure that the test
mainly used in clinical areas (clinical psychology, physiother- can measure motor performance dispositions defined in the
apy). Motor skill tests assess gross and fine motor skills and are course curricula. In this respect, this new instrument differs
mainly used as a diagnostic instruments for motor deficits in explicitly from already available tests like the TGMD-2.
early childhood to six-year-old children. The gross motor skills Furthermore, the evaluation of effects in PE in school
are frequently assessed through test items on object control, requires the measurement of learnable elements of motor
locomotion, and balance. Fine motor skills are measured activity, as these identify the core of teaching procedures in
through test items on hand coordination (e.g., the handling of PE. Thus, it is necessary to consider situation-specific and
paper and scissors). Also, in some areas these instruments are context-dependent characteristics of PE as well as prior
used to assess the motor development of preschool children experiences of pupils. Another respect in which our instru-
(Motor test for four- to six-year-old children [MOT 4–6]; ment differs from most existing tests is the design of the test
Zimmer & Volkamer, 1987). In their review, Cools et al. items, which are related explicitly to PE and vary in difficulty
(2009) give a good overview of the skill-oriented test instru- depending on the age of the pupils. The test instrument can
ments currently in use. therefore be adapted to the performance levels of different
The main field of application (motor deficits, motor devel- class levels. Finally, we attempt to design a test battery that
opment) implies that the test item difficulties should be rela- can be implemented in instructional practice and conducted
tively low, which in turn leads to poor differentiation of by teachers. The goal is to develop a valid test instrument
performance levels in regular school samples. This, plus the whose tasks ensure a simple and practical evaluation.
missing link to sport contexts (with the exception of Test of
Gross Motor Development [TGMD–2]; Ulrich, 1985) and the
syllabus in PE, leads to a restricted curricular validity. METHOD
The MOBAQ approach (Herrmann, Leyener et al., sub-
mitted; Kurz et al., 2008) was developed from a sport pedago- The test instrument presented in this study includes a total of
gical tradition and refers to open motor dispositions (Bös, 2001; 10 test items (Table 1, in detail Herrmann & Seelig, 2014) and
Roth & Willimczik, 1999). The term basic qualifications serves as the collection of basic motor competencies
expresses the fact that these motor performance dispositions (“Motorische Basiskompetenzen” [MOBAK]) of primary
represent the basal requirements for learning specific skills and school pupils in grade 1 (pupils aged 6–7 years). This
 82 HERRMANN, SEELIG, AND GERLACH

TABLE 1
MOBAK Test Items

(C1) Balancing: Being able to balance.

Level 1 The pupil crosses an upside-down long bench which is lying on the ground fluently without leaving the bench; no follow steps.
Level 2 The pupil crosses an upside-down long bench which is lying on a springboard fluently without leaving the bench; no follow steps.
(C2) Rolls: Being able to roll forward.
Level 1 The pupil performs a forward roll down an inclined mat fluently and is able to land in a standing position on his feet.
Level 2 The pupil performs a forward roll down an even mat track fluently and is able to land in a standing position on his feet.
(C3) Throwing: Being able to throw at a target.
Level 1 The pupil throws six juggling balls from a 1.5 m distance at a target of 130 cm in height using an overhead cast.
Level 2 The pupil throws six juggling balls from a 2 m distance at a target of 130 cm in height using an overhead cast.
(C4) Catching: Being able to catch.
Level 1 The test leader drops a basketball from 2 m to the ground so that the ball bounces at least 1.30 m above the ground. The pupil catches the
ball after the reversal point.
Level 2 Ditto, with tennis ball
(A1) Jumping: Being able to jump rhythmically.
Level 1 The pupil jumps across four carpet tiles (40 × 40 cm, distance of 40 cm) fluently. Legs are closed between the tiles. Legs are straddled
beneath the tiles.
Level 2 The pupil jumps across four carpet tiles (40 × 40 cm, distance of 40 cm) fluently. One-legged between the tiles. Legs are straddled beneath
Downloaded by [188.103.189.14] at 11:54 19 May 2015

the tiles.
(A2) Bouncing: Being able to bounce in a controlled manner.
Level 1 The pupil bounces a small basketball at least four times and turns on his own axis once.
Level 2 The pupil bounces a small basketball through a marked corridor (5 × 1 m) without losing the ball.
(A3) Running & Jumping: Being able to vary running and jumping.
Level 1 The pupil runs 3 m to two longitudinal mats, jumps forward with two legs, runs 3 m and jumps with two legs on two mats again, and runs
3 m to arrive at the finish line.
Level 2 Ditto; the pupil jumps on one leg on the mats.
(A4) Pushing: Being able to push up.
Level 1 The pupil pushes on a chin-high vault box and holds his body weight for about two seconds with outstretched arms.
Level 2 The pupil climbs onto a chin-high vault box and stands on the box on his feet
(A5) Dribbling: Being able to dribble.
Level 1 The pupil guides the ball on his foot while turning around.
Level 2 The pupil dribbles a football with at least five ball contacts through a marked corridor (5 × 1 m) without losing the ball.
(A6) Sidestepping: Being able to walk sideways and backwards.
Level 1 The pupil walks 3 m forward twice along a ground marking and walks back backwards without leaving the ground marking.
Level 2 The pupil moves 3 m sideways along a ground marking on the right side twice and moves back on the left side without leaving the ground
marking.

Note. C = common items: data collected in every class; A = additional items: incomplete test plan.

instrument has to differentiate through a broad performance developing the test items may be understood as a design step
level and thus has to include test items with different difficulty for maintaining curricular and content validity.
levels. Furthermore, the test items have to be embedded into In our case, the focus of the item construction was on the
the syllabus (curricular validity), assigned to a latent factor connection to curricular conditions which define minimal
(factorial validity), and measure the same constructs for boys standards in PE at the elementary schools of the City of
and girls (test fairness). The final test instrument is based on a Zurich (Sportamt Zürich, 2013). Besides the aforementioned
larger pool of test items developed and discussed in several aspects of test fairness, further criteria for construction of the
expert discussions. In order to ensure the content validity of the final item set were curricular and factorial validity as well as
test items, we invited seven experts1 on early-childhood (phy- the considerations of feasibility and age-appropriate item
sical) education, sports motor skills, psychometrics, and sports design. In the first stage of test development, we designed
pedagogy to participate in the discussion. This process of each of the 10 test items with two different levels of difficulty.

Design
1
The experts recruit from University of Basel (Switzerland, Department
of Sport, Exercise and Health), University of Potsdam (Germany, Faculty of This study was conducted in parallel to the annual motor-
Human Sciences) and University of Teacher Education (Northwestern performance inventory of motor abilities (“Sportmotorische
Switzerland). Bestandsaufnahme motorischer Fähigkeiten” [SMBA]; c.f.
 ASSESSMENT OF BASIC MOTOR COMPETENCIES 83

Tomatis, Siegenthaler, & Krebs, 2011). The data collection Only in the descriptive evaluation is the actual non-esti-
took place in November 2013 at 10 schools in the Canton of mated data reported.
Zurich. The test included a total of 16 classes with 317 pupils Every analysis was carried out with the help of a prob-
(age: M = 7.04 years, SD = .37, Min. = 6.14, Max. = 8.36; girls: abilistic test model (also: item response theory [IRT]), which
n = 173, boys: n = 144) in the first grade. provides the essential advantage that the measurements are
While the SMBA data collection was created as a full survey sample-independent (Rost, 2003) and the item parameter
in the Canton of Zurich, we used a subsample of 10% which could be estimated independently from the person parameter.
represented the geographical distribution (city, country, etc.) for The manifest (observable) results of the test items are related
this study. The study was conducted in class during regular to the latent factors with a probability mass function. Hence,
teaching time. A double period of 90 minutes was available to one can make assumptions about the probability with which a
seven classes (N = 138 pupils) and a usual period of 45 min. pupil will pass a test item (Bühner, 2011; Strauß, Büsch, &
was available to nine classes (N = 179 pupils). In the classes Tenenbaum, 2006; Tenenbaum, Strauß, & Büsch, 2007).
with a double period, we were able to implement all 10 test The parameter estimation of all estimations was made
items with both levels of difficulty. Due to the time limit in the with the robust weighted least square (WLSMV) estimation
other classes, we were not able to implement a complete test- algorithm implemented in Mplus. A two-parameter normal
ing. Thus, we developed an incomplete test plan in order to try ogive item response theory model (2PNO-IRT) was
out all 10 test items in these classes. Four test items were embedded into an expanded structural equation model. On
selected which were collected in every class (“common the basis of the assumption that the continuously distributed
Downloaded by [188.103.189.14] at 11:54 19 May 2015

items,” see Table 1). The selection of common items is justified latent variables rely on the not-continuously distributed
by the comparability of contextual and methodical comparabil- observable variables, the relations between the manifest
ity as well as the coverage of the broadest possible task spec- test items and the latent factor are represented through a
trum. Additionally, we assigned these classes three further test series of probit regression equations (Muthén & Muthén,
items from the remaining six items, which were determined by 2012). This allows the implementation of an exploratory
randomized item permutation completed a priori (c.f. Multi- factor analysis and a confirmatory factor analysis with
Matrix Design; Rost, 2008). Thus, data missing at random were dichotomous and ordinal data. Thus, the benefits of both
created for the three test items not completed, which could then approaches are being exploited: the precise estimation of the
be estimated by means of an imputation procedure (Pokropek, IRT model as well as the evaluation of the factor structure in
2011) and integrated into further statistical modelling. the structure equation model by means of the model-fit
indices.
The data evaluation followed three steps: (1) We deter-
Test Execution
mined the factorial structure with exploratory factor analyses
The standardized test execution and evaluation were and (2) submitted it to confirmatory tests. (3) Subsequently,
ensured through a trained test leader of the University of we examined with the aid of multiple group analysis whether
Basel. One test leader supervised up to three children in the this factorial structure was the same for boys and girls.
circuit operation, where the order for the rotation is deter- For the first step, we divided the total sample (n = 317)
mined by a protocol. randomly into two subsamples. The exploratory factor ana-
After the instruction by the test leader, which included a lyses were based on the data from the first subsample
short description and a one-off demonstration, the pupils (n = 158), while the confirmatory factor analysis was based
had two attempts to complete both levels of difficulty of on the second subsample (n = 159). The exploratory factor
each test item (no trial). The dichotomously scaled MOBAK analysis with the oblique GEOMIN rotation for one to four
test items (0 = failed, 1 = passed) were recorded for each factors was calculated with the 10 performed test items in
level of difficulty and each attempt, with the exception of model 1a. The factor loading and residual variance were
the test items “Throwing” and “Catching.” The pupils had released for estimation and cross loadings were allowed.
six attempts each for these two test tasks. The amount and For the subsequent analysis, we only took items into account
the order of the successful attempts were recorded. that possessed a sufficiently large (≥ .40) significant loading
on only one factor. The same analysis with a shortened test
battery of eight tests items was calculated in model 1b.
Analysis Technique
In the second step, we tested the resulting structure with
The data preparation was carried out in SPSS 21. Further eight test items against restrictive conditions of the confir-
calculations were made with the program Mplus 7.0 matory factor analysis on the basis of the data of the second
(Muthén & Muthén, 2012). By implementing the Full subsample (model 2a). In this confirmatory factor analysis,
Information Maximum Likelihood (FIML) estimation proce- cross loadings were not allowed. Factor loadings and resi-
dures, we were able to consider pupils who showed design- dual variance were estimated freely for each test item.
related missing data (Muthén & Muthén, 2012). All reported Likewise, we repeated this confirmatory factor analysis on
data refer to the complete estimated sample of 317 pupils. the data of the total sample (model 2b). There is a further
 84 HERRMANN, SEELIG, AND GERLACH

restriction in model 2c, where the factor loading of the test Finally, in model 4, we conducted a confirmatory factor
items for each factor is equated within the confirmatory analysis with the covariate gender (MIMIC) based on model
factor analysis (essential tau-equivalent loading; Bühner, 2b in order to be able to estimate the differences in the latent
2011, p. 400). With this, one can examine whether a balance factors between boys and girls. Additionally, we requested
of the factor loading and thus the formation of a factor sum the modification indices (MI > 10) for the direct effect of the
value about a test item of one factor is permitted. covariate gender in the manifest variables of the test items in
Subsequently, the factor sum values of the two factors are order to test for DIF.
calculated (with the non-estimated sample). We conducted a The evaluation of the goodness-of-fit of the models fol-
univariate analysis of variance to examine potential differ- lowed the fit indices suggested by the literature (Bühner,
ences between boys and girls. 2011; Hu & Bentler, 1999; Schermelleh-Engel,
In the third step, the measurement invariance testing was Moosbrugger, & Müller, 2003; Schreiber, Nora, Stage,
examined, which indicates equality of measurement para- Barlow, & King, 2006). The following cut-offs hinted at a
meters. Missing measurement invariances would hint at a very good model adjustment: comparative fit index (CFI) >
presence of a differential item functioning (DIF) and would .95, root mean square error of approximation (RMSEA) >
thus restrict the validity of the results on gender differences .06, standardized root mean residual (SRMR) < .11 (only
(Dimitrov, 2006). In accordance with Geiser (2011) and exploratory factor analysis), weighted root mean square
Dimitrov, (2006), we realized a sequential procedure in residual (WRMR) < .90. As the chi2 value was not inter-
order to examine the measurement invariance of the model pretable without further calculation due to the WLSMV
Downloaded by [188.103.189.14] at 11:54 19 May 2015

concerning gender. This involved taking two steps with estimators, this was not specified in the fit indices. Thus,
increasingly stringent nested models: configural invariance only standardized coefficients were reported continuously
and factorial invariance. due to the better interpretability of the results.
The development of the configural model started with the
specification and the examination of two independent con-
firmatory factor analyses for boys and girls, respectively.
RESULTS
These gender-specific models 3a were called baseline
models.
One of the two levels of difficulty was chosen for each test
We examined the configural invariance in model 3b by
item which had the highest possible differentiation in med-
combining this gender-specific model in a multiple group
ium item difficulty. This was consistently the more difficult
model. This allowed for a model test for boys and girls
second level of the item. The first level of difficulty of the
simultaneously and revealed the baseline comparison values
test items appeared to be too easy for this age group of six-
for the subsequent model 3c with factorial invariance. All
to eight-year-olds. However, for younger age groups, these
parameters were estimated freely. Only the factor structure
test items would probably be appropriate.
was equated between boys and girls (Dimitrov, 2006;
The amount of successful attempts per test item was
Geiser, 2011; Widaman & Reise, 1997). This was done in
added up (0 = no passing, 1 = passed once, 2 = passed
order to make sure that the factor structure (amount and type
twice). The test items “Throwing” and “Catching” were
of latent factors and loadings) was the same for boys and
scaled (0 = 0–2 hits, 1 = 3–4 hits and 2 = 5–6 hits).
girls.
Subsequently, we measured the factorial invariance in
model 3c, which was based on the configural invariance
Descriptive Statistics
and restricted the relation between the items and the latent
factors. We equated additional factor loadings concerning The item difficulties varied in the selected levels of diffi-
the boys and girls in the model by constraining the non- culty of the test items (Table 2). Additionally, the thresholds
standardized factor loading in model 3b above boys and of model 1a were indicated, which made it possible to
girls invariantly (Dimitrov, 2006; Geiser, 2011; Widaman compare the passing probability between the test items.
& Reise, 1997). The tests of gender-specific differences The extent of the threshold parameters for a given person
between the inter-correlation of the latent factors follow parameter determined the probability of passing a test item
the Wald Test of Parameter Constraints. (Bühner, 2011). “Catching” (−2.24/−1.09) was the test item
As model 3b and model 3c were nested, the models with the highest threshold parameter and thus with the high-
could be compared using a difference test. We performed est probability of passing the test item. The test task
this with the Mplus module chi-square difference testing “Throwing” (−0.18/0.96) had the lowest threshold
for WLSMV, as the classical chi2-difference tests was not parameter.
permitted due to the WLSMV estimations (Muthén & On the basis of the distribution and distances of the
Muthén, 2012). Furthermore, we used the model test to threshold parameters, one can assume that the test items
examine whether the intercorrelation between the latent differentiated between the categories as well as between a
variable was significantly different for boys and girls. broad performance level. The threshold parameters
 ASSESSMENT OF BASIC MOTOR COMPETENCIES 85

TABLE 2
Item Difficulties, Samples, and Threshold Parameters

Balancing Rolls Jumping Sidestepping

N % TH N % TH N % TH N % TH

0 17 5.4 67 21.1 97 41.3 35 16.2

1 46 14.5 −1.78 64 20.2 −0.88 78 33.2 −0.24 42 19.4 −0.94
2 254 80.1 −0.91 186 58.7 −0.26 60 25.5 0.58 139 64.4 −0.32
∑ 317 100.0 317 100.0 235 100.0 216 100.0

Dribbling Bouncing Catching Throwing

N % TH N % TH N % TH N % TH

0 46 19.2 96 44.0 4 1.3 143 45.1

1 59 24.6 −0.95 48 22.0 −0.28 45 14.2 −2.24 127 40.1 −0.18
2 135 56.3 −0.33 74 33.9 0.41 268 84.5 −1.09 47 14.8 0.96
∑ 240 100.0 218 100.0 317 100.0 317 100.0

Run & Jumping Pushing Up

Downloaded by [188.103.189.14] at 11:54 19 May 2015

N % TH N % TH

0 34 14.9 49 22.7
1 73 31.6 −1.04 23 10.6 −0.73
2 123 53.5 −0.88 144 66.7 −0.55
∑ 230 100.0 216 100.0

Note: TH = thresholds.

indicated at which increase in a person’s characteristics the solution. The test item “Push Up” is not assigned to any factor
next-highest category could be achieved. due to the moderate, non-significant factor loading of both
The correlation of the test items, age, and gender can be factors. The test item “Running & Jumping” could be assigned
seen in Table 3. The item correlation between the test items to the first factor but demonstrated significant cross loadings
showed minimal to medium values. Gender revealed a lim- on the second factor (Table 5). As a result, we extracted these
ited correlation with the test items “Jumping” and two test items from the test battery and subjected it to a new
“Sidestepping” as well as “Bouncing” and “Throwing.” exploratory factor analysis. The factor analysis (model 1b),
While the girls were better at “Jumping” and conducted again with the shortened test battery of eight test
“Sidestepping,” the boys were better at “Bouncing” and items, extracted a maximum of three factors. On the basis of
“Throwing.” There was also a limited correlation between the course of the eigenvalues ([1] 2.80, [2] 1.71, [3] 1.12 [4]
age and the test items, and thus it seemed that age had no 0.70) and the interpretability regarding the content, we again
relevant influence on test performance. favored the two-factorial solution. The cross loadings were
low and non-significant (Table 5).

Determination of the Factorial Structure

Verification of the Factorial Structure
The exploratory factor analysis in model 1a was calculated
with all 10 test items on the basis of the first subsample. A The cross loadings were low and non-significant (Table 5). The
maximum of three factors could be extracted. We were unable first factor (Figure 1, Table 5) included test items involving
to identify a four-factorial solution. The course of the eigen- movements requiring the entire body. Thus, these factors were
value ([1] 3.44, [2] 1.71, [3] 1.28, [4] 0.87) suggested that two named “locomotion.” The second factor (Figure 2, Table 5)
factors should be extracted by means of the Scree criterion. included test items demanding good skills with a ball and were
The two-factorial solution showed good fit indices, while the thus named “object control.” The two replications of the two-
one-factorial solution showed worse values (Table 4). The factorial structure with confirmatory factor analyses with the
three-factorial solution had high cross loadings at six of the second subsample (model 2a) and the total sample (model 2b)
10 test items, and its content could not be assigned coherently resulted in good model fits (Table 4). The factor loadings in
to the latent factors. Thus, we favored the two-factorial model 2b were within the acceptable range, with values
 86 HERRMANN, SEELIG, AND GERLACH

TABLE 3
Item Correlation of the Complete Estimated Sample (n = 317)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)

(1) Balancing
(2) Rolls .42**
(3) Jumping .48** .27**
(4) Sidestepping .27** .31** .32**
(5) Run & Jumping .32** .11 .26** .26**
(6) Pushing up .41** .47** .20 .28* .21
(7) Dribbling .26* .19* .24* .05 .34** .42**
(8) Bouncing .20 .19* .18 .09 .10 .41** .44**
(9) Catching .15 .31** −.03 .25* .06 .42** .34** .34**
(10) Throwing .08 .10 −.04 −.02 −.04 .18 .35** .45** .33**
Gendera −.07 .18* .27** .24* .17 .04 −.16 −.44** −.16 −.44**
Age −.08 −.01 −.14* −.02 −.10 .08 −.06 .19** .24* .16*
a
Male = 0; Female = 1.
*The correlation has (2-sided) significance at the level of .05l.
**The correlation has (2-sided) significance at the level of .01.
Downloaded by [188.103.189.14] at 11:54 19 May 2015

TABLE 4
Results of the Analysis

Model Analysis Sample CFI RMSEA [CI] SRMR WRMR

1a EFA 1-factor (10 items) SS 1 .80 .076 [.048–.104] .13

EFA 2-factor (10 items) SS 1 .96 .038 [.000–.077] .09
1b EFA 2-factor (8 items) SS 1 .94 .056 [.000–.104] .08
2a CFA 2-factor (8 items) SS 2 .96 .036 [.000–.082] .65
2b CFA 2-factor (8 items) TS .98 .024 [.000–.057] .65
2c CFA tau-equivalent TS .98 .020 [.000–.051] .76
3a CFA baseline model male TS 1.00 .000 [.000–.063] .54
CFA baseline model female TS 1.00 .000 [.000–.056] .55
3b CFA configural invariance TS 1.00 .000 [.000–.052] .88
3c CFA factorial invariance TS 1.00 .005 [.000–.051] .97
4 MIMIC with covariate gender TS .93 .047 [.021–.071] .84

Note. CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval; SRMR = standardized root mean residual;
WRMR = weighted root mean square residual; EFA = Exploratory factor analysis with oblique GEOMIN rotation; CFA = Confirmatory factor analysis;
MIMIC = Confirmatory factor analysis with covariate gender; SS = Subsample (SS1: n = 158; SS2: n = 159); TS = Total sample.

TABLE 5 between .46 and .73. The intercorrelation of the two factors
Factor Structure of the Exploratory Factor Analysis was r = .39 (Figure 1). Thus, our confirmatory factor analyses
confirmed the results of our exploratory factor analyses and
Model 1a Model 1b
thus the two-factorial structure of the MOBAK test items.
F1 F2 F1 F2 We calculated a confirmatory factor analysis with essential
tau-equivalent loadings in model 2c. This model again showed
Balancing .690* .116 .704* .176 a good model fit, slightly better than model 2b (Table 4). The
Rolls .550* −.011 .552* .033
Jumping .786* −.289 .787* −.275
invariant factor loadings of the factors “locomotion” and
Sidestepping .684* .010 .641* .001 “object control” were .59 and .62, respectively. The calculation
Run & Jumping .475* .314* of a sum value within each factor was statistically acceptable.
Pushing up .330 .287 Within each factor, the sum values achieved a maximum of
Bouncing .006 .665* .006 .706* eight points (four items with two points each).
Dribbling .300 .487* .243 .408*
Throwing −.299 .673* .226 .521*
The factor sum values of the two factors are illustrated in
Catching .212 .463* −.294 .643* Table 6, which showed only the cases with complete data. A
slight right-sided distribution for both factors was shown.
Note. *The correlation has (2-sided) significance at the level of .051 One could recognize that the girls were descriptively, but
 FIGURE 1 Confirmatory factor analysis (model 2b).

not significantly (F(1,136) = .039, p = .843, eta2 = .00), control.” Therefore, we tested this with a complete estimated
better at the factor “locomotion,” while the boys performed sample on a latent level in model 4 by calculating a confirma-
significantly (F(1,136) = 14.55, p = .00, eta2 = .10) better at tory factor analysis with the covariate gender (MIMIC). The
Downloaded by [188.103.189.14] at 11:54 19 May 2015

“object control.” model still achieved acceptable fit values (Table 4) and demon-
strated that gender had a significant influence on the latent
factors “locomotion” (r = .20, p = .19) and “object control” (r =
Measurement Invariance Testing −.42, p < .01). On the basis of the requested modification
indices, no modifications concerning the minimal value (MI
The two-factorial solution with eight test items (model 2b)
> 10) were suggested, which could be rated as additional
was calculated in two separate confirmatory factor analyses
evidence that there was no DIF between boys and girls.
(baseline models) for boys (N = 144) and girls (N = 173) in
model 3a. The parameters were released for estimations.
Only the factorial structure was kept equal for boys and
girls. Following the fit values (Table 4), the two models fitted DISCUSSION
well with the data. This showed that the two-factorial struc-
ture and thus the relation between the manifest and latent The exploratory factor analyses with all 10 test items
variables (measuring model) was the same for boys and girls. (Table 1) showed a two-factorial structure with the factors
The two baseline models for boys and girls were com- “locomotion” and “object control.” Due to the high cross
bined to form a multigroup model in model 3b. The model loadings and the low factor loadings, two test items were
fits were excellent and demonstrated that the factor structure gradually extracted from the test instrument, maintaining the
is invariant between boys and girls. factorial structure. We tested the final eight-item test instru-
Afterwards, we tested in multi-group model 3c whether the ment using a confirmatory factor analysis which confirms
relation between the manifest and latent variables (factor load- the explored structure. One further restriction of the con-
ing) for boys and girls was the same (factorial invariance). The firmatory factor analysis using the invariant factor loading
model fit of this model was very good (Table 4). The difference secured the calculation of a factor sum value statistically.
test between model 3b (variant factor loading) and model 3c This enables an easy distribution of the data with easy-to-
(invariant factor loading) showed no significant difference (x2 interpret results in the practical application of the test instru-
[6] = 6.95, p = .33). Furthermore, the unchanged CFI value at ment. This factor sum value differentiates all factor levels
1.00 showed that there was no significant difference between the and demonstrates a distribution with a longer tail to the right
models (Cheung & Rensvold, 2002). This provides the proof in the presented sample. Differences between boys and girls
that the factor loadings between boys and girls were the same, became visible, with boys achieving a better performance in
and thus the measuring model can assumed to be identical. the factor “object control” and girls showing better perfor-
The intercorrelation between the latent variable in the mance in “locomotion.” Thus, the question arose whether
structure model was significantly different for boys and the underlying measuring model had a different factorial
girls (χ2(1) = 4.54, p = .03). There is a stronger correlation structure depending on gender. We tested this using nested
between “locomotion” and “object control” for boys grouping models. The factorial structure did not differentiate
(r = .67, p < .01), than for girls (r = .27, p < .05). significantly between boys and girls. The measuring model
On the basis of the factor sum values, it was shown in was thus invariant in relation to gender, which indicates that
Table 6 that the girls performed better at the factor “locomo- there is no DIF. Therefore, the test instrument is suitable for
tion” and the boys performed better at the factor “object showing the difference between boys and girls.

87
 Downloaded by [188.103.189.14] at 11:54 19 May 2015

88

TABLE 6
Factor Sum Values for Basic Motor Competencies
HERRMANN, SEELIG, AND GERLACH

Total Boys Girls

Loco-motion Object-control Loco-motion Object-control Loco-motion Object-control

N % N % N % N % N % N %

0 Pt. 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0

1 Pt. 6 4.3 4 2.9 2 3.1 1 1.5 4 5.5 3 4.1
2 Pt. 5 3.6 16 11.6 2 3.1 5 7.7 3 4.1 11 15.1
3 Pt. 9 6.5 21 15.2 6 9.2 7 10.8 3 4.1 14 19.2
4 Pt. 14 10.1 28 20.3 5 7.7 9 13.8 9 12.3 19 26.0
5 Pt. 25 18.1 18 13.0 11 16.9 9 13.8 14 19.2 9 12.3
6 Pt. 35 25.4 21 15.2 21 32.3 13 20.0 14 19.2 8 11.0
7 Pt. 30 21.7 22 15.9 15 23.1 15 23.1 15 20.5 7 9.6
8 Pt. 14 10.1 8 5.8 3 4.6 6 9.2 11 15.1 2 2.7
N 138 100.0 138 100.0 65 100.0 65 100.0 73 100.0 73 100.0
Mean 5.48 4.67 5.45 5.29 5.50 4.12
(SD) (1.79) (1.88) (1.65) (1.85) (1.92) (1.75)
Median 6 5 6 6 6 4
 ASSESSMENT OF BASIC MOTOR COMPETENCIES 89

In order to confirm the discriminant validity, the test REFERENCES

battery presented in this study should be distinguished
from and compared to other motor tests. An initial compara- Balz, E. (2009). Fachdidaktische Konzepte update oder: Woran soll sich der
tive study within the presented sample demonstrated that the Schulsport orientieren? [Update on concepts of subject didactics or:
item correlation between the motor ability tests (i.e., stand- Which points of orientation does physical education have?].
Sportpädagogik, 33(1), 25–32.
ing long jump, sprint, c.f. Tomatis et al., 2011) and the Barnett, L. M., Morgan, P. J., van Beurden, E., Ball, K., & Lubans, D. R.
MOBAK test items was moderate (.17 < r < .37). The (2011). A reverse pathway? Actual and perceived skill proficiency and
factors of the motor abilities “coordination” and “speed physical activity. Medicine & Science in Sports & Exercise, 43(5),
strength” explained 55% of the variance in the factor “loco- 898–904. doi:10.1249/MSS.0b013e3181fdfadd
motion” and 33% of the variance in the factor “object Barnett, L. M., van Beurden, E., Morgan, P. J., Brooks, L. O., & Beard, J.
R. (2009). Childhood motor skill proficiency as a predictor of adolescent
control” (Herrmann, Gerlach, Krebs, & Seelig, submitted) physical activity. Journal of Adolescent Health, 44(3), 252–259.
in a latent structural equation model. This suggests that the doi:10.1016/j.jadohealth.2008.07.004
test items displayed different areas of the motor perfor- Bös, K. (Ed.) (2001). Handbuch motorische tests [Handbook of motor
mance, yet they were mutually dependent. tests] (2nd rev. ed.). Göttingen, Germany: Hogrefe.
Bös, K., & Mechling, H. (1983). Dimensionen sportmotorischer Leistungen
Prognostic validity must be seen as the final obstacle. It
[Dimensions of motor performance]. Schorndorf, Germany: Hofmann.
tests whether children or adolescents who reach a high level in Bös, K., Schlenker, L., Büsch, D., Lämmle, L., Müller, H., Oberger, J.,…
the basic motor competencies participate more frequently and Tittlbach, S. (2009). Deutscher Motorik-Test 6–18 [German motor test
longer in sports in their athletic career and whether they show 6–18]. Hamburg, Germany: Czwalina.
Downloaded by [188.103.189.14] at 11:54 19 May 2015

physical and psychosocial health benefits from the participa- Bühner, M. (2011). Einführung in die Test- und Fragebogenkonstruktion
[Introduction to the construction of tests and questionnaires] (3rd ed.).
tion. This can follow only from a longitudinal study, which still
München, Germany: Pearson Studium.
has to be conducted for the presented test instrument. Burton, A. W., & Miller, D. E. (1998). Movement skill assessment.
Nevertheless, the broad data situation from the area of funda- Champaign, IL: Human Kinetics.
mental skills suggests a positive correlation. We were able to Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit
confirm here that a high motor level is closely related to a indexes for testing measurement invariance. Structural Equation
Modeling: A Multidisciplinary Journal, 9, 233–255. doi:10.1207/
better cardio-respiratory endurance performance and activity
S15328007SEM0902_5
level among children, adolescents, and adults (i.e., Barnett Clark, J. E., & Metcalfe, J. S. (2002). The mountain of motor development: A
et al., 2011; Lubans et al., 2010; Stodden & Holfelder, 2013). metaphor. In J. E. Clark & J. H. Humphrey (Eds.), Motor development.
In conclusion, the eight-item test battery used in the Research & reviews (pp. 163–190). Reston, VA: National Association for
present study is suitable for the evaluation of potential effect Sport and Physical Education.
Cools, W., de Martelaer, K., Samaey, C., & Andries, C. (2009). Movement
in PE. The test instrument is beneficial because the testing
skill assessment of typically developing preschool children: A review of
procedure is fast, the test items are easy to evaluate, and the seven movement skill assessment tools. Journal of Sport Science and
results are interpretable without a standard table and statis- Medicine, 8, 154–168.
tical distribution. Participating PE teachers reported a high Dimitrov, D. M. (2006). Comparing groups on latent variables: A structural
acceptance of this battery. equation modeling approach. Work, 26(4), 429–436.
Geiser, C. (2011). Datenanalyse mit Mplus: Eine anwendungsorientierte
The implementation of our test battery for measuring basic
Einführung [Data analysis with Mplus: An application-oriented introduc-
motor competencies fulfills functions on three levels. (1) tion]. Wiesbaden, Germany: VS.
System monitoring: The monitoring of schools leads to infor- Gogoll, A. (2012). Sport- und bewegungskulturelle Kompetenz—Ein
mation on the educational system’s performance. When pro- Modellentwurf für das Fach Sport [Sport and cultural competence—A
vided to educational policy makers, this information can form model for physical education]. In A. C. Roth, E. Balz, J. Frohn, & P.
Neumann (Eds.), Kompetenzorientiert sport unterrichten (pp. 39–52).
the basis of quality improvements; (2) School development:
Aachen, Germany: Shaker.
The reports on pupils’ performance at the individual schools Gundlach, H. (1968). Systembeziehung und körperliche Fähigkeiten und
form the basis for internal reform measures for quality devel- Fertigkeiten [System relations and physical abilities and physical skills].
opment; (3) Competency diagnostics: The individual measure- Theorie und Praxis der Körperkultur, 17, 198–204.
ment of competencies allows teachers to identify groups in Herrmann, C., Gerlach, E., Krebs, A., & Seelig, H. (submitted).
Begründung und Validierung eines Erhebungsinstruments zur
need of special support and initiate special support measures to
Erfassung motorischer Basiskompetenzen in der Primarschule
reduce these inequalities (Klieme et al., 2003). (MOBAK-1) [Explanation and validation of a test instrument for the
Thus, the test battery presented here is not only suitable assessment of basic motor competencies in primary school].
for scientific evaluation but also for the (internal) evaluation Herrmann, C., Leyener, S., Pühse, U., & Gerlach, E. (submitted).
of PE when implemented in normal teaching procedures. On Empirische Überprüfung der Struktur von motorischen
Basisqualifikationen im Schulsport [Empirical investigation of the struc-
the basis of these results, teachers can adapt their teaching
ture of basic motor qualifications in physical education].
content to the current performance level of their pupils. The Herrmann, C., & Seelig, H. (2014). MOBAK-1: Basic motor competencies
teachers do not only receive information about the easily in first grade: Testmanual. Retrieved from http://www.dsbg4public.ch/
determined total value but also concrete information about custom/upload/docs/bx7gklezunvcv4ziuklmf6446rw60cb1251g.pdf
which pupil passed which test item (e.g., 50% of the pupils Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
structure analysis: Conventional criteria versus new alternatives.
cannot perform a forward roll).
 90 HERRMANN, SEELIG, AND GERLACH

Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. Roth, K., & Willimczik, K. (1999). Bewegungswissenschaft [Movement
doi:10.1080/10705519909540118 science]. Reinbek, Germany: Rowohlt.
Klieme, E., Avenarius, H., Blum, W., Döbrich, P., Gruber, H., Prenzel, M., Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2003). Evaluating the
… Vollmer, H. J. (2003). Zur Entwicklung nationaler Bildungsstandards. fit of structural equation models: Tests of significance and descriptive good-
Expertise [The development of national educational standards. ness-of-fit measures. Methods of Psychological Research Online, 8, 23–74.
Expertise]. Frankfurt am Main, Germany: DIPF. Schreiber, J. B., Nora, A., Stage, F. K., Barlow, E. A., & King, J. (2006).
Kurz, D., Fritz, T., & Tscherpel, R. (2008). Der MOBAQ-Ansatz als Konzept für Reporting structural equation modeling and confirmatory factor analysis
Mindeststandards für den Sportunterricht? [Using the MOBAQ approach as a results: A review. The Journal of Educational Research, 99(6), 323–338.
concept for minimum standards for physical education?]. In V. Oesterhelt, J. doi:10.3200/JOER.99.6.323-338
Hofmann, M. Schimanski, M. Scholz, & H. Altenberger (Eds.), Sportamt Zürich. (2013). Minimalstandards der Stadt Zürich für den
Sportpädagogik im Spannungsfeld gesellschaftlicher Erwartungen, wis- Sportunterricht in der Volksschule [Minimal standards of the City of
senschaftlicher Ansprüche und empirischer Befunde (pp. 97–106). Zurich for physical education in primary school]. Retrieved from https://
Hamburg, Germany: Czwalina. www.stadt-zuerich.ch/content/dam/stzh/ssd/Deutsch/Sport/Publikationen
Lubans, D. R., Morgan, P. J., Cliff, D. P., Barnett, L. M., & Okely, A. D. %20und%20Broschueren/11_Unterstuetzung%20Lehrpersonen/Praxispool/
(2010). Fundamental movement skills in children and adolescents: Kompetenztraster_Level_1-4.pdf
Review of associated health benefits. Sports Medicine, 40(12), 1019– Stodden, D., & Holfelder, B. (2013). Kein Kind bleibt zurück: Die Rolle der
1035. doi:10.2165/11536850-000000000-00000 Entwicklung von motorischen Fertigkeiten [No child stays behind: The role of
McKenzie, T. L., & Lounsbery, M. A. F. (2013). Physical education development in motor skills]. Zeitschrift Für Sportpsychologie, 20(1), 10–17.
teacher effectiveness in a public health context. Research Quarterly doi:10.1026/1612-5010/a000088
for Exercise and Sport, 84(4), 419–430. doi:10.1080/ Stodden, D. F., Goodway, J. D., Langendorfer, S. J., Roberton, M. A., Rudisill,
02701367.2013.844025 M. E., Garcia, C., & Garcia, L. E. (2008). A developmental perspective on the
Metztler, M., McKenzie, T. L., Van Der Mars, H., Barrett-Williams, S. L., role of motor skill competence in physical activity: An emergent relationship.
Downloaded by [188.103.189.14] at 11:54 19 May 2015

& Ellis, R. (2013). Health optimizing physical education (HOPE): A new Quest, 60, 290–306. doi:10.1080/00336297.2008.10483582
curriculum for school programs—Part 2: Teacher knowledge and colla- Strauß, B., Büsch, D., & Tenenbaum, G. (2006). Anwendungen probabil-
boration. Journal of Physical Education, Recreation and Dance, 84(5), istischer Modelle in der Sportpsychologie [Applications of probabilistic
25–34. doi:10.1080/07303084.2013.779531 models in sport psychology]. Zeitschrift Für Sportpsychologie, 13(2),
Muthén, L. K., & Muthén, B. O. (2012). Mplus user’s guide: Seventh 68–74. doi:10.1026/1612-5010.13.2.68
edition. Los Angeles, CA: Muthén & Muthén. Tenenbaum, G., Strauß, B., & Büsch, D. (2007). Applications of general-
National Association for Sport and Physical Education. (2011). PE metrics: ized Rasch models in the sport, exercise, and the motor domains. In M.
Assessing national standards 1–6 in secondary school. Reston, VA: V. Davier & C. H. Carstensen (Eds.), Statistics for social and behavioral
AAHPERD. sciences. Multivariate and mixture distribution Rasch models. Extensions
Pokropek, A. (2011). Missing by design: Planned missing-data designs in and applications (pp. 347–356). New York, NY: Springer.
social science. Research & Methods, 20(1), 81–105. Tomatis, L., Siegenthaler, J., & Krebs, A. (2011). SMBA H’11.
Rink, J. E. (2013). Measuring teacher effectiveness in physical education. Sportmotorische Bestandesaufnahme. Motorische Fähigkeiten der
Research Quarterly for Exercise and Sport, 84(4), 407–418. doi:10.1080/ Erstklässler der Stadt Zürich (Herbst 2011) [Inventory of sports-related
02701367.2013.844018 abilities. Motor skills of first-graders in the City of Zurich (Fall 2011)].
Rost, D. H. (2008). Zur Messung von Kompetenzen einer Bildung für Zürich, Switzerland: ETH.
nachhaltige Entwicklung [On the measurement of competencies in the Ulrich, D.A. (1985). Test of gross motor development Austin, Tex: PRO-ED.
education of sustainable development]. In I. Bormann & G. D. Haan Widaman, K. F., & Reise, S. P. (1997). Exploring the measurement invar-
(Eds.), Kompetenzen der Bildung für nachhaltige Entwicklung. iance of psychological instruments: Applications in the substance use
Operationalisierung, Messung, Rahmenbedingungen, Befunde (pp. domain. In K. J. Bryant, M. T. Windle, & S. G. West (Eds.), The science
61–74). Wiesbaden, Germany: VS. of prevention. Methodological advances from alcohol and substance
Rost, J. (2003). Lehrbuch der Testtheorie, Testkonstruktion [Textbook of abuse research (pp. 281–324). Washington, DC: APA.
test theory, test construction]. Bern: Huber. Zimmer, R., & Volkamer, M. (1987). Motoriktest für vier- bis sechsjährige
Roth, K. (1982). Strukturanalyse koordinativer Fähigkeiten [Structure ana- Kinder: MOT 4–6; Manual [Motor test for four- to six-year-old children:
lysis of coordination skills]. Bad Homburg, Germany: Limpert. MOT 4–6; Manual] (2nd ed.). Weinheim, Germany: Beltz.

View publication stats