DEVELOPMENTAL MEDICINE & CHILD NEUROLOGY
ORIGINAL ARTICLE
Validity and reliability of the Movement Assessment Battery for
Children-2 Checklist for children with and without motor impairments
MARINA M SCHOEMAKER 1 | ANUSCHKA S NIEMEIJER 2 | BOUDIEN C T FLAPPER 3 , 4 | BOUWIEN C M
SMITS-ENGELSMAN 5 , 6
1 Centre for Human Movement Sciences, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands. 2 Sint Maartenskliniek, Research
Development and Education, Nijmegen, the Netherlands. 3 Department of Paediatrics, Effatha Guyot Centre, Groningen, the Netherlands. 4 University Medical Center Groningen,
Groningen, the Netherlands. 5 Avans+, University for Professionals, Breda, the Netherlands. 6 Faculty of Kinesiology and Rehabilitation Sciences, KU Leuven, Belgium.
Correspondence to Dr Marina M Schoemaker, University Medical Center Groningen, Sector F, Center for Human Movement Sciences, PO Box 196, 9700 AD Groningen, the Netherlands. E-mail:
[email protected]
PUBLICATION DATA
AIM The aim of this study was to investigate the validity and reliability of the Movement
Accepted for publication 4th October 2011.
Published online 9th February 2012.
METHOD Teachers completed the Checklist for 383 children (age range 5–8y; mean age 6y 9mo;
ABBREVIATIONS
DCD
DCDQ'07
MABC-2
TMS
TTS
Developmental coordination
disorder
Developmental Disorder
Coordination
Questionnaire 2007
Movement Assessment Battery for
Children-2
Total motor score
Total test score
Assessment Battery for Children-2 Checklist (MABC-2).
190 males; 193 females) and the parents of 130 of these children completed the Developmental
Disorder Coordination Questionnaire 2007 (DCDQ’07). All children were assessed with the MABC2 Test. The internal consistency of the 30 items of the Checklist was determined to measure reliability. Construct validity was investigated using factor analysis and discriminative validity was
assessed by comparing the scores of children with and without movement difficulties. Concurrent
validity was measured by calculating correlations between the Checklist, Test, and the DCDQ’07.
Incremental validity was assessed to determine whether the Checklist was a better predictor of
motor impairment than the DCDQ’07. Sensitivity and specificity were investigated using the
MABC-2 Test as reference standard (cut-off 15th centile).
RESULTS The Checklist items measure the same construct. Six factors were obtained after factor
analysis. This implies that a broad range of functional activities can be assessed with the Checklist,
which renders the Checklist useful for assessing criterion B of the diagnostic criteria for DCD. The
mean Checklist scores for children with and without motor impairments significantly differed
(p<0.001). The scores for the Checklist ⁄ Test and DCDQ’07 were significantly correlated (rS=)0.38
and p<0.001, and rS=)0.36 and p<0.001, respectively). The Checklist better predicted motor impairment than the DCDQ’07. Overall, the sensitivity was low (41%) and the specificity was acceptable
(88%).
INTERPRETATION The Checklist meets standards for validity and reliability.
According to the diagnostic criteria in the DSM-IV-TR,1 children with developmental coordination disorder (DCD) have a
marked impairment in motor coordination which hampers the
performance of academic and daily-life motor skills, without
evidence of a medical disorder or intellectual impairment.1
Early identification of the condition is recommended,2–4 but
correct identification is impeded by the lack of a universally
accepted criterion standard for DCD.5,6 To solve this problem, a multiple assessment approach has been advocated to
assess whether the inclusion criteria for DCD have been
met.7–9 Motor questionnaires can be used to investigate
whether the performance of daily-life motor skills at home or
at school is affected (criterion B). A motor test can also be
administered to assess the extent and severity of the movement
difficulties (criterion A).
Several questionnaires have been developed to gather information about functional motor performance from parents
(Developmental Disorder Coordination Questionnaire 2007
[DCDQ’07];10 Children Activity Scales for Parents11), teachers (Teacher Estimation of Activity Form;12 Motor Observation Questionnaire for Teachers;13 Children Activity Scales
for Teachers;11 Movement Assessment Battery for Children
Checklist [MABC Checklist]14,15), or from the children themselves (Children’s Self-perceptions of Adequacy in and Predilection for Physical Activity16). Few motor tests are available
to assess objectively the severity and extent of children’s movement difficulties (e.g. MABC-215 and the Bruininks–Oseretsky
Test of Motor Proficiency); the MABC is most commonly
used in research.17 Despite the popularity of the MABC Test,
its counterpart, the MABC Checklist, has been used far less
often in clinical studies. One of the reasons for this is probably
its length (48 items), which has rendered it too laborious to
administer.
Recently, both the Test and the Checklist of the MABC
have been revised.15 The length of the Checklist has been
reduced to 30 items, and the content of some of the items has
368 DOI: 10.1111/j.1469-8749.2012.04226.x
ª The Authors. Developmental Medicine & Child Neurology ª 2012 Mac Keith Press
been changed. Although the reliability and validity of the original version of the Checklist have been investigated,7,8,18 a
limitation of the revised version is the lack of research on its
reliability and validity.19 Therefore, the aim of the present
study was to investigate these psychometric properties. The
internal consistency of the items of the Checklist was investigated as a measure of reliability. Concurrent validity was
examined in two ways: first, by conducting correlations
between the Checklist, the Test, and the DCDQ’07 as a
parental questionnaire and, second, by investigating whether
children with scores in the clinical range on the MABC-2
Checklist also fail on the Test. In order to diagnose a child
with DCD, both criteria A and B of the diagnostic criteria for
DCD have to be met. According to the manual,15 one of the
functions of the Checklist is to assist in diagnosing DCD as it
provides a view of how a child performs everyday movement
tasks at home or at school (criterion B). The MABC-2 Test is
often used to ascertain whether criterion A has been met (i.e.
there is a marked impairment in motor coordination). Consequently, the classification agreement between Test and Checklist is interesting from both a clinical and a psychometric point
of view. Lastly, the extent to which the Checklist has any
added value with respect to existing questionnaires was
addressed by investigating whether the Checklist increased
detection of motor impairment compared with the
DCDQ’07,10 a popular parental questionnaire used to identify
children with DCD.
METHOD
Participants
The Dutch standardization study of the MABC-2 includes a
Dutch and a Flemish (Dutch-speaking part of Belgium) sample. The Dutch sample is representative of the Dutch population regarding sociodemographic status (ethnicity, parental
educational level, distribution of geographic region; see the
manual of the Dutch MABC-220). For the standardization
study, teachers were asked to randomly select five Dutch children aged between 3 and 16 years from their school class,
which gave a total of 825 students. Of these students, all 500
students aged 5 to 8 years were selected for the present study.
The teachers of these students were asked to fill out the
Checklist. The participating teachers had on average 12 years’
teaching experience (range 1–40y). Of the 447 Checklists that
were returned, 64 had to be excluded because more than three
items in one section were missing, giving a final sample of 383
children (86%). Five items were difficult to fill out for teachers; presumably those skills are not regularly observed in a
school situation: (A1.5, ‘Pours liquid from one container to
another’; B22, ‘Hits a moving ball with a bat’; B25, ‘Participates in a team game’; B31, ‘Rides a bicycle without stabilizers’; B3,3 ‘Maintains balance in water’). All 383 children were
assessed with the MABC-2 Test, which has eight items, and
gives a total test score ranging from 8 to 152. About one-third
of the parents of the 383 participating children (randomly
selected) were asked to fill out the DCDQ’07 (n=130). Either
the father or the mother could fill out the questionnaire. The
educational level of the father was measured on a 5-point scale
What this paper adds
• This is the first study to investigate the reliability and validity of the MABC-2
Checklist.
• The Checklist can discriminate between children with and without motor
impairment.
• The Checklist is useful for assessing criterion B of the diagnostic criteria for
developmental disorder coordination.
as a predictor of socio-economic status (1=only primary school
education; 5=university education). The median educational
level of the father (3.0) was representative of the educational
level in the Dutch population.
Informed consent was obtained from parents of the children. The study was approved by the local medical ethics
committee.
Procedure
Sixteen therapists (mean age 25y) with at least 4 years’ experience in the administration of MABC-1 were trained to administer the revised version of the MABC-2 Test. The MABC-2
Test (age bands 1 and 2) was administered in a quiet room at
each child’s school.
Instruments
MABC-2 Checklist
The Dutch translation of the Checklist with accompanying
Dutch norms was used in this study. A forward–back translation procedure was adopted to translate the Checklist into
Dutch. The Checklist is designed to identify children with
motor difficulties in the age range 5 to 11 years. It has a motor
and a non-motor part. The motor part contains 30 items
divided into two sections. Section A measures movement in a
static and ⁄ or predictable environment and section B measures
movement in a dynamic and ⁄ or unpredictable environment.
The non-motor part was outside the scope of the present
study. Each section is subdivided into three parts, each containing five items. Section A measures self-care skills, classroom skills, and physical education ⁄ recreational skills; section
B measures self-care ⁄ recreational skills, ball skills, and physical
education ⁄ recreational skills. For each item, teachers have to
rate the motor competence of a child on a 4-point scale
(0=very well; 3=not close). The total motor score (TMS) is the
sum of the 30-item scores; the higher the TMS, the poorer
the performance. According to the manual of the MABC-2,15
children with scores at or above the 95th centile are highly
likely to have a motor impairment in daily-life, children with
scores between the 85th and 94th centile are ‘at risk’ of having
a motor impairment, and children with scores up to the 85th
centile have no detectable motor impairment.
Movement ABC-2 Test
The aim of the Test is to classify children of 3 to 16 years
11 months of age according to degree of motor impairment.
There are three age-related item sets, each consisting of
eight items measuring manual dexterity (three items), aiming
and catching (two items), and balance (three items). Following the procedure described in the MABC-2 manual,15 item
scores can be transformed into standard scores (mean=10;
Validity of the MABC-2 Checklist Marina M Schoemaker et al. 369
SD=3).15 The total test score (TTS) is the sum of the eightitem standard scores (range 8–152). Standard scores are also
provided in the manual for the TTS (mean=10; SD=3).15
Test–retest reliability of the TTS is good (0.80).15,20,21 The
previous version of this Test had acceptable validity, but so
far information about the validity of the revised version is
limited.19 Dutch norms of the Test and Checklist were used
in this study.20
DCDQ'07
The DCDQ’07 is a parent questionnaire developed to identify
motor problems in children aged from 5 to 15 years.10 The
questionnaire contains 15 items regarding motor coordination. For each item, parents have to compare the degree of
coordination of their child with other children of the same
age, and rate this on a 5-point scale. By adding the scores for
each item, a total DCDQ score can be calculated. The higher
the score, the better the motor performance level. The internal consistency of the questionnaire is high (Cronbach’s
alpha=0.89).
Data analysis
When fewer than four questions on the Checklist contained
missing values, missing values were replaced as instructed in
the manual of the MABC-2.15 If an item was missing from any
of the subsections within either section A or B, then the
remaining four items in that section were reviewed. If the
child’s scores were 0 or 1, than a score of ‘1’ was assigned for
the missing item, if the scores were 2 or 3, than a score of ‘2’
was assigned for the missing item. In cases where scores were
mixed, a score of ‘1’ was assigned.15
Reliability
Cronbach’s alpha was calculated to determine the degree of
consistency among the 30 motor items of the Checklist.
Construct validity
In order to explore which items of the Checklist cluster
together in a meaningful way, a factor analysis (principal components analysis with varimax rotation) was carried out on the
raw item scores. Varimax rotation was used to maximize the
loading of each item onto one of the extracted factors whilst
minimizing the loading on all other factors. This simplifies
the interpretation as it results in a small number of items loading highly onto each factor.22 The factor structure is achieved
by looking for items that correlate highly with other items
(loading >0.40) but do not correlate with other items outside
that group. A two-way analysis of variance was carried out to
investigate the effects of age (4) and sex (2) on the TMS of the
Checklist. As the assumption of homogeneity of variance was
violated, the analysis was performed on the logarithmic transformation of the TMS scores.
Discriminant validity
In the present study, the sample was split into two groups, a
group with standard scores up to 7 on the MABC-2 Test
(including children with significant motor impairment and
370 Developmental Medicine & Child Neurology 2012, 54: 368–375
those at risk for motor impairment; henceforth called the
group with motor impairments) and a typically developing
group with total standard scores above 7 on the MABC-2
Test. A total standard score of 7 is associated as closely as
possible with the 15th centile cut-off. The 15th centile was
chosen as the cut-off point as it is recommended in the European guidelines for assessment and intervention for DCD as
the cut-off point for motor impairment when using a motor
test. A logistic regression analysis was performed in order to
predict the classification of children with and without motor
impairments, with TMS scores as the predictor variable. A
receiver operating characteristic curve was composed in order
to provide a graphical picture of the ability of the Checklist
to discriminate between children with and without motor
impairment based upon their MABC-2 test scores. In addition, the mean TMS of the group with motor impairments
was compared with the mean TMS of the typically developing group (analysis of variance). The number of items per
age group for which significant differences were obtained
(t-test) between the groups was also calculated in order to
investigate the discriminative ability of the Checklist across
age groups.
Concurrent validity
The degree of concurrent validity between the Test and the
Checklist and between the DCDQ’07 and the Checklist was
examined using Spearman’s rank order correlation on MABC2 TSS, DCDQ’07 total scores, and the Checklist’s TMS. The
concurrent validity was also investigated by calculating the
percentage of agreement between the Checklist and the Test
in classifying children as those with motor impairment
(impaired and at-risk groups) and those without motor impairment, and by calculating the sensitivity and specificity of the
Checklist. Sensitivity refers to the percentage of children with
problems according to the Test that is correctly detected by
the Checklist (80% is preferable).23 Specificity refers to the
percentage of children without problems who are correctly
identified (90% is preferable). Cohen’s kappa was calculated
to evaluate the proportion of agreements that is actually
observed between the Checklist and the Test in classifying
children as those with and those without motor impairment
after adjusting for the proportion of agreements that take place
by chance.24
Incremental validity
Logistic regression analysis was conducted to determine
whether TMS scores of the Checklist or DCDQ’07 total
scores were better able to predict the presence or absence of
motor impairment on the MABC-2 Test. Logistic regression
was chosen as the outcome was a categorical variable (children
with motor impairments [standard scores £7 on MABC-2
Test] vs typically developing children [standard scores >7]). In
step 1, the DCDQ’07 total scores were entered as predictor
variables with age and sex as possible confounders. In step 2,
the Checklist’s TMS were added. The assumptions for regression analysis (i.e. no multicollinearity) were checked beforehand. Alpha was set at 0.05.
Table I: Mean (SD) total motor scores for the Movement Assessment Battery for Children-2 (MABC-2) Checklist and mean (SD) total standard
scores for the MABC-2 Test for age and sex separately
Age
(y)
n
5
97
6
94
7
84
8
108
Total
383
Sex (n)
Checklist total
motor score
MABC-2 total
standard score
Males (47)
Females (50)
Males (51)
Females (43)
Males (48)
Females (36)
Males (44)
Females (64)
Males (190)
Females (193)
11.9 (11.2)
11.8 (10.6)
11.5 (10.6)
7.3 (6.8)
10.6 (11.6)
3.2 (4.6)
4.8 (6.3)
2.2 (3.2)
9.7 (11.2)
6.0 (7.8)
9.2 (3.4)
9.3 (3.2)
9.0 (3.2)
10.5 (3.1)
8.7 (3.5)
11.5 (3.4)
9.2 (3.4)
11.1 (3.2)
9.0 (3.4)
10.6 (3.4)
RESULTS
Table I provides an overview of the number of children across
age and sex, and the mean and SD of scores on the Checklist
and Test. Twenty-five per cent of the Checklists had one, two,
or three items missing, which in the majority of cases included
the item ‘Maintains balance in water among other children
(B3.3)’, as teachers in the Netherlands generally do not
observe children in water.
Construct validity
Cronbach’s alpha was 0.94 for all 30 items together, suggesting that all items measure the same construct.
Factor structure of the motor items of the Checklist
Six factors were extracted with an eigenvalue of less than 1,
which together explain 69% of the variance (see Table II for
an overview of the six factors, their content, and factor loading). The first factor explained most of the variance (20%) and
was formed by items requiring gross motor coordination during self-care or classroom activities. Factor 2 was formed by
items involving ball skills, factor 3 by items measuring recreational skills, factor 4 by items measuring fine motor skills, factor 5 by items that measure the ability to keep rhythm, and
factor 6 by items that involve dynamic balance. Three items
did not load on any factor: A1.5, ‘Pours liquid from one container to another’; A3.3, ‘Throws a beanbag or ball so that
another stationary child can catch it’; and B3.1, ‘Rides a bicycle without stabilisers’. Teachers noted that two of these items
(A1.5 and B3.1) were difficult to fill out as they can generally
not be observed in a school setting (missing data: 19% for
both items).
A significant effect of age (F3,375=21.329; p<0.001) and sex
(F1,375=14.056; p<0.001) was obtained for the TMS, as well as
a significant interaction between them (F3,375=2.651; p=0.049).
As the effect of age was significant, post-hoc analyses were carried out for age using the Bonferroni procedure. The results
revealed that the mean TMS of 8-year-old children differed
significantly from the mean TMS of all other age groups. The
mean TMS of the 6- and 7-year-old children did not differ
significantly from each other, but did differ significantly from
those of the 5- and 8-year-old children. The TMS were equal
for males and females at 5 years of age, but after that age the
mean TMSs of males declined less than those of females.
Discriminant validity
The TMS was a significant predictor of motor impairment ⁄ non-motor impairment (B [standard error, SE]=)0.082
[0.015]; p<0.001 [odds ratio 0.92; CI 0.90–0.95]). Figure 1
shows the ability of the Checklist to discriminate between
children with and without motor impairment based upon
their MABC-2 Test scores. The results of the analysis of
variance revealed that, for each age group, the mean TMSs
of children in the group with motor impairments significantly differed from those of children in the typically developing group (see Table III). The number of items
discriminating significantly between the group with motor
impairments and the typically developing group for 5-, 6-,
7-, and 8-year-old children was respectively 21, 16, 22, and
11 out of 30 items.
Concurrent validity
A significant correlation (Spearman’s rho) was obtained
between the TMS and the TTS (rS=)0.38; p<0.001; n=383),
and between the total scores of the DCDQ’07 and the TMS
(rS=)0.36; p<0.001; n=130). The groups used for calculating
the correlation between the TMS and the TTS, on the one
hand, and the TMS and the DCDQ’07, on the other hand,
did not differ significantly in age (6y 9mo and 6y 8mo,
respectively; p=0.12). The correlation between the TMS and
the TTS for the smaller group (n=130) was comparable to
that of the whole group (rS=)0.35; p<0.001). The percentage
agreement between the Checklist and Test was 80%
(307 ⁄ 383) (see Table IV). A kappa value of 0.28 was
obtained, which is considered to be fair. Sensitivity was 41%
(26 ⁄ 64) and specificity 88% (281 ⁄ 319) across all age groups if
the 15th centile was used as a cut-off for the Test and the
Checklist. Overall, 16.7% (n=64) of the children performed
below the 15th centile on both instruments, and 6.8% and
5.7% performed below the 5th centile on the Test and the
Checklist, respectively.
Incremental validity
In step 1 of the logistic regression analysis, DCDQ’07 total
scores significantly predicted the outcome on the MABC-2
Test (R2=4.6%). When the Checklist TMSs were entered into
the analysis in step 2, they significantly predicted the outcome
on the MABC-2 Test (R2=7.3%), whereas the DCDQ’07 total
scores no longer significantly added to the predicted outcome
(see Table V).25
DISCUSSION
The aim of this study was to investigate the psychometric
properties of the motor part of the MABC-2 Checklist in a
sample of children almost twice as large as the sample of 203
5- to 8-year-old children described in the MABC-2 manual.15
To start with, the internal consistency of the items of the
Checklist was very good, which implies that each question
addressed an aspect of functional motor performance.
Validity of the MABC-2 Checklist Marina M Schoemaker et al. 371
Table II: Factor loadings of the six factors found after factor analysis on the 30 items of the motor part of the Movement Assessment Battery for Children-2
Checklist
Item
Maintains balance while standing to pull on clothing (A1.1)
Puts on clothing over the head (A1.2)
Washes and dries hands (A1.4)
Walks around classroom avoiding objects and persons
(A2.4)
Transports objects without dropping them (A2.5)
Crosses the playground avoiding collision with
objects ⁄ persons (A3.5)
Maintains balance when frequent adjustments are required
(B1.1)
Moves around a busy classroom collecting objects (B1.2)
Carries a tray around a room avoiding moving persons
(B1.3)
Crosses the playground avoiding collision with moving
persons (B3.5)
Catches a ball with two hands (B2.1)
Hits a moving ball with a bat (B2.2)
Throws a ball on the move to another child (B2.3)
Continually bounces a large ball (B2.4)
Participates in a team game using skills of throwing and
catching (B2.5)
Uses stationary playground equipment (A3.4)
Participates in chasing games (B3.2)
Maintains balance in water among other children (B3.3)
Uses non-stationary playground equipment (B3.4)
Fastens buttons (A1.3)
Manipulates small objects (A2.1)
Forms letters using a pen (A2.2)
Uses scissors to cut paper (A2.3)
Keeps time to a musical beat by clapping hands (B1.4)
Moves body in time with music (B1.5)
Jumps keeping two feet together on take off and landing
(A3.1)
Hops on either foot (A3.2)
Eigenvalue
Amount of variance explained (%)
Factor 1,
gross motor
coordination
skills
Factor 2,
ball skills
Factor 3,
recreational
skills
Factor 4,
fine motor
skills
0.57a
0.57
0.51
0.79
0.18
0.09
)0.03
0.08
0.28
0.42
0.47
)0.05
0.24
0.22
0.04
0.22
0.30
0.24
0.32
0.08
0.23
)0.16
)0.09
0.15
0.82
0.79
0.10
0.14
0.05
0.09
0.11
0.10
0.14
0.02
0.14
0.25
0.63
0.07
0.26
0.16
0.26
0.26
0.71
0.64
0.14
0.32
0.32
0.11
0.16
0.19
0.22
0.34
0.15
)0.06
0.77
0.15
0.28
0.15
)0.03
0.24
0.18
0.06
0.18
0.18
0.12
0.53
0.87
0.85
0.80
0.78
0.23
0.13
0.03
0.12
0.15
0.35
0.07
0.09
0.18
)0.01
0.20
0.07
0.13
0.17
0.09
0.34
0.11
0.13
0.17
)0.03
0.21
0.04
0.16
0.23
0.33
0.29
0.19
0.27
0.30
0.25
0.33
0.04
0.22
0.30
0.07
0.20
0.01
0.15
0.03
0.28
0.28
0.20
0.69
0.76
0.56
0.79
0.32
)0.04
)0.01
0.09
0.16
0.14
0.33
)0.01
0.10
0.16
0.05
0.60
0.53
0.77
0.70
0.19
0.18
0.11
0.13
0.01
)0.06
0.14
)0.09
0.47
0.18
0.25
0.75
0.76
0.09
0.23
)0.03
0.26
0.11
)0.16
0.22
0.18
0.23
0.10
0.13
0.65
0.39
6.059
20
0.24
4.075
16
0.08
3.336
11
0.26
2.659
8
Factor 5,
rhythmic
skills
0.17
2.327
8
Factor 6,
dynamic
balance
0.58
1.815
6
a
Factor loadings above 0.50 are printed in bold.
Exploration of the factor structure of the Checklist revealed a
six-factor solution, implying that the Checklist measures a
broad range of functional motor abilities. This is important
for an instrument designed to assess functional motor impairment in children at risk for DCD considering the heterogeneous nature of the disorder.
In comparable questionnaires, such as the Motor Observation Questionnaire for Teachers and the DCDQ’07, gross
motor skills were found to explain most of the variance.
The factor solution of the Checklist is in line with these
findings, as the three factors explaining most of the variance
all consist of items requiring gross motor coordination. The
finding that gross motor skills are covered by three factors
instead of one is probably due to the larger number of items
in the Checklist that cover a particular aspect of gross
motor performance, such as ball skills or recreational skills,
which cluster together. In addition, similar to the Motor
Observation Questionnaire for Teachers and the DCDQ’07,
372 Developmental Medicine & Child Neurology 2012, 54: 368–375
a factor covering fine motor skills was found. Factor analysis
on the Checklist also revealed two additional factors (rhythmic skills and dynamic balance skills) which did not appear
as separate factors in the DCDQ’07 or Motor Observation
Questionnaire for Teachers, either because none of the
items referred to these skills or because they were covered
by only one item.
Significant but moderate correlations were obtained
between the Checklist and the Test scores ()0.38), and
between the Checklist and the DCDQ’07 ()0.36), which
implies that these instruments assess both overlapping but also
slightly distinct motor skills. The obtained correlation
between the Checklist and the Test is lower than that found in
the UK standardization sample for 5- to 12-year-old children
()0.55) (Barnett A, personal communication 2010) and also
lower than the correlations reported in previous populationbased studies between the original MABC Test and Checklist,
which varied between 0.44 and 0.51.8,18
1.0
ROC Curve
Sensitivity
0.8
0.6
0.4
AUC=0.673
0.2
0.0
0.0
0.2
0.4
0.6
1 - Specificity
0.8
1.0
Figure 1: Receiver operating characteristic (ROC) curve for the Movement Assessment Battery for Children-2 (MABC-2) Checklist in relation to
MABC-2 Test performance. AUC, area under the curve.
The results regarding incremental validity imply that scores
on the Checklist are a better predictor of outcome on the
MABC (performance above or below 15th centile) than scores
on the DCDQ’07. However, the fact that the DCDQ’07 was
no longer a significant predictor of motor impairment when
the Checklist was added as a predictor in the regression analysis actually indicates that the questionnaires are highly correlated. Consequently, it is too premature to conclude that
teachers are better able to rate motor performance than parents. In addition, the content of the questions in both questionnaires might also be responsible for the results regarding
incremental validity. A study in which both questionnaires are
filled out by parents and teachers might shed more light on
this issue.
The Checklist was found to be able to discriminate between
the group with motor impairments and the typically developing group across all ages. As expected, a developmental trend
was found. Older children in both groups were able to master
more functional skills than younger children according to their
teachers, which was reflected in a decrease in the number of
items discriminating between the two groups (age 5y: 21 out
of 30 items; age 8y: 11 out of 30 items). Males obtained poorer
scores on the Checklist than females.
An important question is whether children with functional
impairments as measured with the Checklist (Criterion B) also
have motor impairments as measured with the Test (Criterion
A). Overall, the classification agreement between the Checklist
and the Test was 80%, which is comparable to the agreement
found for the UK standardization sample (78%; Barnett A,
personal communication 2010). Also, sensitivity and specificity
rates were comparable in the Dutch and UK samples (sensitivity Dutch sample, 41% vs 50% in the UK sample; specificity
Dutch sample, 88% vs 84% in the UK sample). This implies
that only 41% of the children with motor impairments on the
Test have functional motor impairments at school as measured with the Checklist. Low rates for either sensitivity or
specificity are a common finding in population-based samples,
irrespective of which questionnaire measuring functional
motor impairment is used and irrespective of whether parents
or teachers are asked to rate motor performance.3,12,18 A practical implication of these findings is that the Checklist cannot
be recommended for use for population-based screening
owing to its low sensitivity, which is in agreement with the
advice in the manual of the MABC-2 and with the recommendations of the European Academy of Childhood Disability
(EACD) for assessment and intervention for children with
DCD regarding DCD questionnaires in general.15,26
Several explanations can be given for the findings regarding
sensitivity and specificity. First, in the present study, a random
sample of mainstream school children was included. As a consequence, the number of children with motor impairment was
rather small. In addition, performance on the MABC-2 Test
was only used as a criterion measure. For future studies, it
might be informative to include a clinical sample of children
who meet all four diagnostic criteria for DCD, as the sensitivity and specificity of an instrument can be more reliably determined in a sample in which 50% of included children have
DCD.27 Second, school teachers are generally not formally
trained to observe and rate motor development, which may
have affected the results. Finally, the assessment of motor dysfunction is complicated by the multifactorial nature of motor
capacity. To provide a reliable picture of a child’s motor skills,
measurement of a large number of skills is required, especially
in a heterogeneous condition such as DCD.28 However, so
Table III: Mean (SD) total motor scores of the Movement Assessment Battery for Children-2 Checklist across age for the group with motor impairment ⁄ at
risk for motor impairment and the typically developing group
Age (y)
Motor impairment ⁄ at risk for
motor impairment, n (TTS £15th%)
Typically developing
group, n (TTS >15th%)
95% CI on difference
between group means
p-value
5
6
7
8
17, 22.3 (15.0)
13, 18.0 (12.3)
15, 15.3 (15.3)
19, 6.3 (7.3)
80, 9.6 (8.3)
81, 8.2 (10.0)
69, 5.3 (7.1)
89, 2.6 (3.9)
8.00–17.10
3.75–15.85
4.95–15.05
1.39–6.01
<0.001
0.002
<0.001
0.002
TTS, total test score; CI, confidence intervals.
Validity of the MABC-2 Checklist Marina M Schoemaker et al. 373
Table IV: Agreement between Movement Assessment Battery for
Table V: Results of logistic regression analysis with Developmental Disor-
Children-2 (MABC-2) Test and Checklist in classifying a child as motor
der Coordination Questionnaire 2007 (DCDQ'07) total scores and Move-
impaired or not motor impaired
ment Assessment Battery for Children-2 (MABC-2) Checklist scores as
predictor variables and motor impairment ⁄ no motor impairment as criteMABC-2 Checklist
MABC-2 Test
Motor impairment
No motor impairment
Total
rion variable
Motor
impairment
No motor
impairment
Total
26
38
64
42
277
319
68
315
383
Probability that both instruments rate child as motor impaired:
(68 ⁄ 383)·(64 ⁄ 383)=3%. Probability that both instruments rate
child as not motor impaired: (319 ⁄ 383)·(315 ⁄ 383)=68%.
Kappa=(PA)PC) ⁄ (1)PC)=(79.1)71) ⁄ (1)71)=28%. PA, observed per
cent agreement=(26+277) ⁄ 383=79%. PC, per cent agreement
expected from chance=71% (68%+3%).
far, an empirically determined ‘criterion standard’ is lacking
for the identification of DCD.27 In the absence of a true criterion standard, conclusions regarding the sensitivity and specificity of the Checklist need to be treated with caution.
LIMITATIONS OF THE STUDY
Only the internal consistency of the motor items could be
studied as an aspect of reliability. No data regarding test–retest
reliability were available. It would be worthwhile including
those aspects of reliability in future studies. In addition, the
psychometric properties of the Checklist were assessed in a
mainstream school sample in which the number of children
with movement difficulties was small. This may have influenced our findings.
95% CI for odds ratio
B (SE)
p-value
Lower
Odds ratio
Upper
Step 1
Age
Sex
DCDQ’07
)0.16 (0.22)
0.57 (0.48)
0.05 (0.02)
0.46
0.24
0.009
0.56
0.69
1.01
0.85
1.76
1.05
1.30
4.53
1.09
Step 2
Age
Sex
DCDQ’07
MABC-2
Checklist
)0.38 (0.24)
0.33 (0.51)
0.03 (0.02)
)0.07 (0.03)
0.12
0.52
0.19
0.01
0.43
0.52
0.99
0.88
0.69
1.39
1.03
0.93
1.10
3.75
1.07
0.98
R2=0.046 (Hosmer and Lemeshow)29, 0.052 (Cox and Snell)30 for step 1,
R2=0.073 (Hosmer and Lemeshow), 0.081 (Cox and Snell) for step 2
(p=0.01).
CLINICAL IMPLICATIONS
The Checklist is a valid instrument that provides information
about the performance of a child in a broad range of activities
of daily living. Therefore, it is a useful tool for assessing Criterion B of the diagnostic criteria for DCD.
ACKNOWLEDGEMENTS
The authors would like to thank Dr Anna Barnett for her valuable comments on the manuscript.
REFERENCES
1. American Psychiatric Association. DSM-IV-TR. Diagnostic
and Statistical Manual of Mental Disorders, 4th edn, text
revision. Washington, DC: American Psychiatric Association, 2000.
2. Jongmans MJ. Early identification of children with Developmental Coordination Disorder. In: Sugden DA, Chambers
ME, editors. Children with Developmental Coordination
Disorder. London: Whurr Publishers, 2005, 155–67.
3. Schoemaker MM, Flapper B, Verheij NP, Wilson BN, Rein-
disorder in Singapore. Dev Med Child Neurol 1996; 38:
1099–105.
14. Henderson SE, Sugden DA. The Movement Assessment Bat-
Psychometric properties of the movement ABC checklist as a
tery for Children. San Antonio, TX: The Psychological Cor-
screening instrument for children with Developmental Coordination Disorder. Br J Educ Psychol 2003; 73: 425–41.
9. Sugden DA. Leeds Consensus Statement: Economic Science
Research Council Seminar Series. Cardiff: Dyscovery Trust,
2006.
10. Wilson BN, Crawford SG, Green D, Roberts G, Aylott A,
mental Coordination Disorder Questionnaire as a screening
Kaplan BJ. Psychometric properties of the revised Develop-
instrument. Dev Med Child Neurol 2006; 48: 668–73.
mental Motor Coordination Questionnaire. Phys Occup Ther
Pediatr 2009; 29: 182–202.
BE, Hay JA. Evaluating the CSAPPA sub-scales as potential
11. Rosenblum S. The development and standardization of the
screening for developmental coordination disorder. Arch Dis
Children Activity Scales (ChAS-P ⁄ T) for the early identifica-
Child 2007; 92: 987–91.
tion of children with developmental coordination disorders.
5. Henderson SE, Barnett AL. The classification of specific
motor coordination disorders in children: some problems to
be solved. Hum Mov Sci 1998; 17: 449–69.
6. Larkin D, Rose E. Assessment of developmental coordination disorder. In: Sugden DA, Chambers ME, editors. Children with Developmental Coordination Disorder. London:
Whurr Publishers, 2005: 135–54.
7. Wright HC, Sugden DA. A two-step procedure for the iden-
27: 190–9.
8. Schoemaker MM, Smits-Engelsman BCM, Jongmans MJ.
ders-Messelink HA, De Kloet A. Evaluation of the Develop-
4. Cairney J, Veldhuizen S, Kurdyak P, Missiuna C, Faught
developmental coordination disorder. Hum Mov Sci 2008;
Child Care Health Dev 2006; 32: 619–32.
12. Faught BE, Cairney J, Hay J, Veldhuizen S, Missiuna C, Spi-
poration, 1992.
15. Henderson SE, Sugden DA, Barnett AL. Movement Assessment Battery for Children – 2 Examiner’s Manual. London:
Harcourt Assessment, 2007.
16. Hay JA, Hawes R, Faught BE. Evaluation of a screening
instrument for developmental coordination disorder. J Adolesc Health 2004; 34: 308–13.
17. Geuze RH, Jongmans MJ, Schoemaker MM, Smits-Engelsman BCM. Clinical and research diagnostic criteria for developmental coordination disorder: a review and discussion.
Hum Mov Sci 2001; 20: 7–47.
18. Junaid K, Harris SR, Carswell A. Teacher’s use of the MABC
Checklist to identify children with motor difficulties. Pediatr
Phys Ther 2000; 12: 158–63.
ronello CA. Screening for motor coordination challenges in
19. Brown T, Lalor A. The Movement Assessment Battery for
children using teacher ratings of physical ability and activity.
Children-Second Edition (MABC-2): a review and a critique.
Hum Mov Sci 2008; 27: 177–89.
Phys Occup Ther Pediatr 2009; 29: 86–103.
13. Schoemaker MM, Flapper BCT, Reinders-Messelink HA,
20. Smits-Engelsman BCM. Movement Assessment Battery for
De Kloet A. Validity of the motor observation questionnaire
Children-2. Dutch Standardization. Pearson Education:
for teachers as a screening instrument for children at risk for
Amsterdam, 2010.
tification of children with developmental co-ordination
374 Developmental Medicine & Child Neurology 2012, 54: 368–375
21. Smits-Engelsman BC, Niemeijer AS, van Waelvelde H. Is
25. Hunsley J, Meyer GJ. The incremental validity of psycholog-
the Movement Assessment Battery for Children-2nd edition
ical testing and assessment: conceptual, methodological, and
a reliable instrument to measure motor performance in
statistical issues. Psychol Assessment 2003; 15: 446–55.
3 year old children? Res Dev Disabil 2011; 32: 1370–7.
22. Field A. Discovering Statistics using SPSS. London: Sage
Publications, 2005.
23. American Psychological Association. Standards for Educational and Psychological Tests. Washington, DC: American
Psychological Association, 1985.
24. Cohen J. A coefficient of agreement for nominal scales. Educ
with Developmental Coordination Disorder? Br J Occup
Ther 2005; 68: 2–10.
28. Cools W, De Martelaer K, Samaey C, Andries C. Movement
26. Blank R, Smits-Engelsman B, Polatajko H, Wilson P. Euro-
skill assessment of typically developing preschool children: a
pean Academy for Childhood Disability (EACD): recom-
review of seven movement skill assessment tools. J Sports
mendations on the definition, diagnosis and intervention of
developmental coordination disorder (long version). Dev
Med Child Neurol 2012; 54: 54–93.
Sci Med 2008; 8: 154–68.
29. Hosmer DW, Lemeshow S. Applied Logistic Regression.
New York: Wiley, 1989.
27. Green D, Bishop T, Wilson BN, et al. Is questionnaire-based
screening part of the solution to waiting lists for children
30. Cox DR, Snell DJ. The Analysis of Binary Data, 2nd edn.
London: Chapman & Hall, 1989.
Psychol Meas 1960; 20: 37–47.
Validity of the MABC-2 Checklist Marina M Schoemaker et al. 375