The development of fundamental movement skills (FMS) is an essential requirement for children to successfully participate in sport, exercise, and physical activity (5). Hence, several studies demonstrate the reciprocal effect between FMS and physical activity (eg, 30). The development of FMS allows children to participate in various physical activities and to develop a higher level of movement skills, which is in the long term associated with participation in an active lifestyle (9, 34, 70).
With the goal of making children’s lives more physically active, research has shown that large differences in the motor competence level exist between different age, sex, and weight status groups (9, 34, 42). The latter is particularly relevant as obesity has also become a global concern in children. Recent data revealed that worldwide 39 million children under the age of 5 were overweight or obese in 2020, an increase of nearly 6 million since 2000 (82). Children with excessive weight tend to have a lower performance level of FMS (7, 42). Vice versa, children with a low level of FMS more likely become overweight or obese (61). In preschoolers with overweight or obesity, it is therefore essential to measure their developmental status of FMS because children with a low level of FMS are less likely to participate in sports and physical education (9, 33, 60, 70).
The compulsory education setting offers an ideal opportunity to potentially reach most children and to counteract the current decline in physical activity levels observed during childhood (30). Consequently, an important goal of physical education is the promotion of early FMS development (5, 30). To achieve this objective by adapting targeted strategies, a most valid assessment of the FMS provides relevant information. Thus, it is hardly surprising that various test batteries have emerged since the last century (10, 51).
These motor test batteries vary in their factor structure of FMS, that is, in the number and type of distinct categories or constructs as well as the respective test items (54). Most assessment tools are used to evaluate FMS by locomotor (such as jumping, hopping, or running) and object (such as catching or striking a ball) control. However, recent research has shown that balance is an important aspect of FMS (63) that is mostly overlooked (81) or not represented as a separate factor in assessments of kindergarten children. Instead, balance is often seen as a part of locomotor or object control. Balance and stability rely on vestibular sensors in the inner ear, vision, and proprioceptors throughout the body, which feed sensory information to an internal model or sensorimotor map (18). In infancy and early childhood, this model depends heavily on visual input (53). Around ages 5–6, the reliance shifts from vision to somatosensory and vestibular inputs. Balance and locomotion skills develop in parallel but do not mature in an adult-like manner until adolescence (53). Compared with balance and locomotion, object manipulation skills tend to develop later because of the more challenging visual-motor requirements of these tasks (24). To take into account that balance control is fundamentally important for skilled movements (17), our SALzburg Together against Obesity (SALTO) test instrument treats balance (including a static and dynamic balance task) as a separate factor in addition to locomotion and object manipulation (cf. 30, 59).
The established assessments further vary in the fundamental type of measurement (ie, product- or process-oriented). Product-oriented tests measure the outcome of a movement (eg, time, distance, speed), whereas process-oriented assessments focus on the quality of the movement pattern or whether it meets specific execution criteria (44). Several test batteries exist for either product- or process-oriented assessments (22, 37). Even though a few hybrid instruments for collecting and evaluating product- and process-oriented data are available for older age groups (eg, 45, 73), such assessments are still lacking for kindergarten children. This limits the comparison of results as process- and product-oriented measurements assess different aspects (55) and provide distinct insights into the motor competence level across skills and age (55, 81). Furthermore, many test batteries are proposed in school settings for children with a normal fitness level and especially with normal body mass. With sex- and age-specific differences being already well documented (9, 50), body weight may be another strong moderator of FMS (4). A higher body weight is associated with biomechanical limitations as children with overweight or obesity are more likely to suffer from movement-restricting orthopedic problems and have greater difficulty in physical activities of daily living, which, in turn, can lead to poorer FMS (7, 42, 50).
Breaking these limitations by applying an assessment tool for all preschoolers that accounts for the meaningful dimensions of FMS as well as combines a product- and process-oriented approach would lead to a better understanding of FMS in this age group and would further help to promote motor competence and movement skills in children. Therefore, the aim of the study was to evaluate both a product- and a process-oriented assessment approach representing 3 predefined constructs of FMS in preschoolers. In detail, the purpose of the study was (1) to test a model of locomotion, object manipulation, and balance as latent variables (multitrait) considering the method-specific effects of product- and process-oriented assessments (multimethod) and (2) to evaluate the model invariance across subgroups of age, weight, and sex in a preschool cohort. By confirming the structural validity and measurement invariance, kindergarten teachers obtain a feasible test battery that helps enhance skill development across all 3 building blocks of FMS by physical education programs based on individual test results. The test battery can also be applied in academic studies to identify and compare trends in the development of FMS as well as to examine the relevance and effectiveness of intervention programs. This, in turn, is critical for gaining support from stakeholders, such as parents, schools, and funding bodies. From a developmental perspective, the kindergarten age is a transitional period between the initial and mature stages of FMS development, in which the progress depends partly on growth and maturation and largely on the opportunity for motor practice, encouragement, high-quality instructions, and environmental factors (21). During this period, there is an acceleration of the cortical maturation and the cortex becomes more organized (19, 58), which makes it particularly receptive to learning and refining basic motor skills. Consequently, it seems sensible to focus on kindergarten children when creating a suitable instrument for assessing FMS.
Methods
Study Design
The development of the test battery for the assessment of FMS in preschool children was part of the longitudinal-sequential study SALzburg Together against Obesity (SALTO) (59). From all childcare centers (n = 350) of the federal state of Salzburg (Austria), 21 services—representing every district—were included in the study. Participation was contingent upon the unanimous agreement of the entire preschool team (management, teachers, and assistants) to implement the SALTO activities for a period of 2–3 years. The annual baseline assessment of FMS in 36- to 72-month-old children took place from November 2014 to June 2018, with half of the children tested in autumn and the other half in spring. Overall data were collected for a sample of 736 children (356 girls) with a mean age of 4.53 (0.85) years (for demographic characteristics, see Table 1). All testers completed a standardized workshop on test instruction and execution 2 times. Children were tested only if a written informed consent from the legal guardian was given. The study was approved by the ethics committee of the University of Salzburg (Austria) (EK-GZ 25) and conducted with the permission of the office for early childcare centers of the province Salzburg.
Demographic Characteristics
| Sample size | Age, y | B height, m | B weight, kg | BMI, kg/m2 | BMI-SDS | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | % | M | SD | M | SD | M | SD | M | SD | M | SD | |
| Male | 380 | 51.6 | 5.03 | 0.82 | 1.11 | 0.07 | 19.6 | 3.4 | 15.8 | 1.6 | 0.27 | 1.04 |
| Female | 356 | 48.4 | 5.02 | 0.76 | 1.10 | 0.07 | 18.9 | 3.4 | 15.5 | 1.5 | 0.07 | 0.99 |
| Young | 357 | 48.5 | 4.34 | 0.41 | 1.06 | 0.05 | 17.6 | 2.7 | 15.7 | 1.5 | 0.23 | 1.01 |
| Old | 379 | 51.5 | 5.67 | 0.44 | 1.15 | 0.05 | 20.8 | 3.2 | 15.6 | 1.6 | 0.12 | 1.03 |
| Normal | 599 | 81.4 | 5.04 | 0.79 | 1.10 | 0.07 | 18.5 | 2.6 | 15.1 | 0.9 | −0.17 | 0.71 |
| Overweight | 137 | 18.6 | 4.93 | 0.77 | 1.12 | 0.08 | 22.7 | 3.9 | 18.0 | 1.5 | 1.69 | 0.77 |
| Overall | 736 | 100 | 5.02 | 0.79 | 1.11 | 0.07 | 19.3 | 3.4 | 15.6 | 1.6 | 0.17 | 1.02 |
Note: BMI-SDS, body mass index standard deviation score as recommended by the World Health Organization (Ending Childhood Obesity 2016, Child Growth Standards: Methods and Development 2016).
Anthropometrics
Before testing FMS, each child was measured barefoot in light clothing for body mass and body height with a standardized balance scale and an attached stadiometer (Seca 899 + 225) to the nearest unit (0.1 kg, 0.5 cm). According the recommended procedure of the World Health Organization, the body mass index (BMI, kg·m−2) of the children was standardized for age and sex (BMI-standard deviation score) by using a web-based tool (PC PAL WebCalc version 2.0, 2016-07-15, pcpalsoftware.eu). To determine the percentage of the children with normal weight and with overweight or obesity, we utilized the sex-specific weight-for-height values for children under 5 years as well as the sex-specific BMI-for-age values for 5- to 7-year-old children as recommended by the World Health Organization Child Growth Standards (80). Children with a BMI-for-age value between +1 SD ≤ z < +2 SD were classified as overweight (n = 97; 13%) and with a z score ≥ +2 SD as obese (n = 40; 5%). Children with a BMI-for-age value between −1 SD < z < +1 SD were categorized as normal weight (n = 517; 70%) and with a z score ≤ −1 SD as underweight (n = 82; 11%) (80). For statistical analyses, children classified as normal or underweight were combined into one group, as were those classified as overweight or obese.
Testing Fundamental Movement Skills
The SALTO test battery consisted of movement tasks with different item difficulty assigned to 3 motor components (latent factors): Locomotion, Object Manipulation, and Balance. All motor tests were selected based on the age of the children, motor development, and the awareness level of the motor test in German-speaking countries for comparing the data (72, 75, 82). Furthermore, recognizing the quality of the completion of a motor task (I can jump—I do jump) as introduced by Burton and Miller (13), we selected process-oriented in addition to product-oriented measures as recommended by others (75, 79). The following 10 motor tests (11 items) were selected to measure the level of FMS by a process- and a product-oriented assessment method (see Table 2).
Overview of the FMS Tests
| Type of assessment | Motor component | FMS | Test criteria, outcome** | Author |
|---|---|---|---|---|
| Process-oriented* | Object manipulation | Catching | Score 0–6 | Ulrich (TGMD-2 and -3) |
| Kicking | Score 0–8 | |||
| Throwing | Score 0–8 | |||
| Locomotion | Hopping | Score 0–8 | Ulrich (TGMD-2 and -3) | |
| Broad jump 1 (horizontal jumping) | Score 0–8 | |||
| Product-oriented | Locomotion | Broad jump 2 (horizontal jumping) | Distance in centimeters | Bös et al (11, KMS) |
| Forward roll | Score 0–6 (concerning the end-position of the roll) | SALTO, acc. Haywood and Getchell (33) | ||
| Shuttle run | Time in seconds for running a distance of 4 × 4 m | Krombholz (41, MoTB) | ||
| Lateral jumping | Total number of lateral jumps in 2 trials of 15 s | Bös et al (11, KMS) | ||
| Balance | Balance beam (walking on a beam) | Average number of steps in 2 trials on a 3-m long, 3-cm broad, and 1-cm high beam | Krombholz (41, MoTB) | |
| One-leg stand | Time in seconds of standing on the preferred leg (max. 30 s) | SALTO, mod. to Bös et al (11, KMS) |
Abbreviations: FMS, fundamental movement skills; KMS, Karlsruher Motorik-Screening; MoTB, Motorik-Testbatterie; SALTO, SALzburg Together against Obesity; TGMD, Test of Gross Motor Development.
*Test criterion dichotomously scaled as 0 = not performed correctly and 1 = performed correctly. The movement quality of each task is described by 3 or 4 motor characteristics. Two attempts were counted. Accordingly, a total score of 0 to 6 or 8 per motor task can be achieved by the child. **High values depict high-performance level.
To measure the skill level of object manipulation, that is, how a child can handle a ball, we selected 3 motor tasks (skills) of the Test of Gross Motor Development (TGMD-2 and -3) by Ulrich (74, 75): 2-hand catch, kick of stationary ball, and overhand throw. All 3 skills were assessed by rating the quality of the performance level dichotomously, that is, if the movement criteria were presented correctly (score 1) or not (score 0). Each motor task is described by 3 or 4 criteria, such as “Child’s hands are positioned in front of the body with the elbows flexed” (75, p. 19), and was completed twice (2 trials). This assessment is called process-oriented because it evaluates how a movement is performed and describes qualitative movement patterns. For further details on criteria prescription, the reader is referred to Ulrich (74, 75).
Locomotion was assessed process-oriented by testing Hopping 4 times on the preferred leg (score 0–8 [75]), process- and product-oriented by jumping horizontally from a standing position as far as possible (Broad Jump 1: score 0–8 [75]; Broad Jump 2: distance in centimeters [72]) and only product-oriented by testing the Forward Roll (level of end-position, score 0–6 [33]), the Shuttle Run (time in seconds for a distance of 4 × 4 m [41]), and Lateral Jumping (number of jumps in 2 × 15 s [72]).
The balance performance was assessed product-oriented both by a dynamic test of walking forward on a 3-m long/3-cm broad/5-cm high beam (2 trials on the Balance Beam, total number of steps over 2 × 3 m with a maximum of 8 steps [41]) and a static test of standing on the preferred leg as long as possible (One-leg Stand, time in seconds [72]).
All tests were explained in a child-friendly manner and demonstrated, and each child conducted a practice trial. The tests were provided in a course of 10 stations and passed through in groups of 4 to 6 children. The testing for one kindergarten group of 15 to 25 children was completed in 4 morning hours (8.00 a.m. to 12.00 noon).
The SALTO test team was trained in 3 steps: First, testers had to read the test manual through self-study. This was followed by a 90-minute classroom session with a demonstration of the test procedure and a subsequent role-play testing. Finally, the introduction was concluded with a clarification of questions about the test procedure and feedback from a sports scientist.
The test results were reported to both the kindergarten teachers and the children’s parents. The teachers received feedback on the test performance—along with reference values—for the motor tasks (individual and group) and on the BMI percentile of the individual child. The individual test results were also communicated to each child’s parents.
Statistical Analyses
In a first step, the data set was cleaned of unrealistic values, and subsequently outliers were removed using 2 conservative numerical detection techniques. With this, only extreme outliers were eliminated if the value of a test item exceeded the Grubbs criterion (α = .05; 2-sided) and lay outside the Tukey’s fences of 3 times the interquartile range. Applied iteratively, the procedure resulted in the removal of 7 values across all test items.
For the confirmatory factor analysis of the single-indicator multitrait–multimethod model of the SALTO test battery, in a second step, an overall 3 factor model was specified in Mplus 8 (52) with the DIFFTEST option for the statistical between-group test of the model fits. The FMS Hopping, Broad Jump 1 (process), Broad Jump 2 (product), Forward Roll, Shuttle Run, and Lateral Jumping served as indicators for the factor Locomotion. The FMS Catching, Kicking, and Throwing formed the factor Object Manipulation. Finally, the FMS Balance Beam and One-leg Stand were included into the factor Balance. These factors were modeled as latent variables. The skills Broad Jump 2, Forward Roll, Shuttle Run, Lateral Jumping, Balance Beam, and One-leg Stand assessed the product of the skill execution, and Hopping, Broad Jump 1, Catching, Kicking, and Throwing evaluated the quality of the movement skill. Thus, 2 different assessment methods were used. Therefore, a latent method factor (23) was introduced to address this multimethod design, including all variables that addressed product assessments (see Figure 1). Above and beyond this measurement variance captured by the method factor, the same items shared additional measurement characteristics. Therefore, following the correlated uniqueness approach (47), the correlations of the residuals of the following manifest variables (uniquenesses) were estimated: (1) all continuous variables, (2) all categorical variables using 7 categories, and (3) all categorical variables using 9 categories. Finally, because Broad Jump 1 and Broad Jump 2 were assessed based on the same task, the correlation of the uniqueness of each was estimated as well. Due to the employment of both continuous and categorical indicators, the weighted least square mean and variance adjusted estimator, implemented in Mplus, was used for model estimation. Locomotion was scaled according to Hopping (9 categories: 0–8), Object Manipulation according to Catching (7 categories: 0–6), and Balance according to One-leg Stand (3 categories: 1–3).
Overall SALzburg Together against Obesity multitrait–multimethod model with standardized (StdYX) factor loadings and intercorrelations.
Citation: Pediatric Exercise Science 2025; 10.1123/pes.2024-0076
To arrive at a model that allows for between-group comparisons, in a third step, the overall model was tested on measurement invariance across age (young: 3–4 y, n = 357; old: 5–6 y, n = 379), BMI (normal weight: n = 599; overweight: n = 137), and sex subgroups (girls: n = 356; boys: n = 380). We chose the hierarchical procedure suggested by Cheung and Rensvold (16). First, the configural model in which the latent variables are specified is based on the same set of identical items across groups. If the model fit is sufficient, the factor loadings are then set equal across groups, testing for metric or weak invariance. If the model fit does not change, these restrictions can be accepted. As suggested by Rutkowski and Svetina (64), we used the change in comparative fit index (ΔCFI) and root-mean-square-error of approximation (ΔRMSEA) as comparison criteria with a value of ΔCFI ≤ −.02 and a value of ΔRMSEA ≤ .03 indicating that the null hypothesis of metric invariance should not be rejected. Finally, scalar or strong invariance was tested by restricting the item intercepts (continuous variables) or thresholds (categorical variables) as being equal across group, using ΔCFI ≤ −.01 and ΔRMSEA ≤ .01 as comparison criteria, following Rutkowski and Svetina (64).
In all models, missing data were handled by the Full Information Maximum Likelihood algorithm (2). Thus, cases with missing data were not deleted, but all model parameters were estimated based on the cases with complete data and the (conditional) missing values under the missing at random assumption. Compared with listwise deletion, this procedure of dealing with missing data avoids common disadvantages, such as losing statistical power and biased parameter estimation (29, 65).
Two-sided t tests were used to test for differences between groups in the test items and in the latent variables based on the factor scores of the multigroup model. In addition to the frequentist t test and Cohen d as effect size, Bayesian analyses were conducted using JASP 0.16 (jasp-stats.org) with the default Cauchy prior distribution centered at zero and a scaling parameter of 0.707 to quantify relative evidence for equivalence. For significance testing of group differences in intercorrelations, Fisher r-to-z transformation was applied with Cohen q as effect size (62).
Results
Overall Model
The path diagram of the SALTO multitrait–multimethod model is illustrated in Figure 1. The overall model yielded good model fit indices: χ2 (23, n = 736) = 38.9, P = .020, CFI = .994, Tucker–Lewis Fit Index = .985, RMSEA = .031. The means and SDs of the test items and latent variables are depicted in Table 3. The intercorrelations are reported in Table 4. Because the intercorrelations between Locomotion and Object Manipulation and between Locomotion and Balance turned out to be relatively high, 2-factor models were specified in each case to test whether both factors can be combined into a single factor. The fit of the model that combined Locomotion and Object Manipulation proved to be sufficient: χ2 (26, n = 736) = 222.3, P < .001, CFI = .923, Tucker–Lewis Fit Index = .838, RMSEA = .101, but significantly worse than the fit obtained by the 3-factor model: Δχ2=100.3, Δdf = 3, P < .001. Similarly, the fit of the model that combined Locomotion and Balance proved to be good: χ2 (26, n = 736) = 66.80, P < .001, CFI = .981, Tucker–Lewis Fit Index = .966, RMSEA = .046, but also significantly worse than the fit obtained by the 3-factor model: Δχ2 = 17.39, Δdf = 3, P < .001. Therefore, we chose the 3-factor model for further analyses.
Mean (SD) of the SALTO Test Items and Latent Variables, P Values, BF10, and Effect Sizes (Cohen d) of the Inferential Analyses (2-Sided Frequentist and Bayesian t Test for Independent Samples) Between Subgroups
| Age | BMI | Sex | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Young | Old | P | BF10 | d | Normal | Overweight | P | BF10 | d | Male | Female | P | BF10 | d | |
| Process-oriented | |||||||||||||||
| Catching, pt | 3.7 (1.5) | 4.4 (1.4) | <.001 | >1000 | 0.490 | 4.1 (1.5) | 3.8 (1.4) | .032 | 0.770 | −0.192 | 4.1 (1.5) | 4.1 (1.5) | .781 | 0.086 | 0.021 |
| Kicking, pt | 4.4 (1.9) | 4.8 (1.9) | .004 | 4.9 | 0.214 | 4.6 (1.9) | 4.7 (1.8) | .549 | 0.122 | 0.053 | 5.2 (1.8) | 4.0 (1.8) | .000 | >1000 | −0.672 |
| Throwing, pt | 4.0 (2.1) | 4.6 (2.0) | <.001 | 290.3 | 0.304 | 4.3 (2.1) | 4.0 (2.2) | .123 | 0.361 | −0.151 | 4.8 (2.1) | 3.8 (1.9) | .000 | >1000 | −0.490 |
| Hopping, pt | 3.8 (2.4) | 5.2 (2.0) | <.001 | >1000 | 0.640 | 4.6 (2.3 | 4.2 (2.3) | .051 | 0.688 | −0.187 | 4.4 (2.3) | 4.7 (2.3) | .124 | 0.265 | 0.114 |
| Broad jump 1, pt | 4.7 (2.2) | 5.4 (2.0) | <.001 | >1000 | 0.335 | 5.1 (2.1) | 4.7 (2.2) | .043 | 0.811 | −0.194 | 5.1 (2.1) | 5.1 (2.2) | .894 | 0.083 | 0.010 |
| Product-oriented | |||||||||||||||
| Broad jump 2, cm | 79.2 (19.6) | 99.4 (18.9) | <.001 | >1000 | 1.053 | 90.2 (21.9) | 87.3 (20.4) | .131 | 0.287 | −0.136 | 92.3 (22.0) | 86.8 (21.0) | .001 | 26.2 | −0.254 |
| Forward roll, pt | 2.1 (1.4) | 2.9 (1.5) | <.001 | >1000 | 0.580 | 2.6 (1.6) | 2.3 (1.5) | .072 | 0.505 | −0.172 | 2.5 (1.5) | 2.5 (1.6) | .809 | 0.086 | 0.018 |
| Shuttle run, s | 11.2 (2.0) | 9.5 (1.3) | <.001 | >1000 | −1.044 | 10.2 (1.9) | 10.5 (1.9) | .156 | 0.292 | 0.138 | 10.4 (2.0) | 10.2 (1.7) | .249 | 0.159 | −0.086 |
| Lateral jumping, n | 23.2 (8.1) | 35.1 (11.7) | <.001 | >1000 | 1.167 | 29.6 (11.8) | 28.9 (11.5) | .575 | 0.123 | −0.053 | 29.4 (12.1) | 29.5 (11.4) | .858 | 0.085 | 0.013 |
| Balance beam, n | 2.0 (1.2) | 3.3 (1.9) | <.001 | >1000 | 0.769 | 2.8 (1.7) | 2.3 (1.5) | .001 | 7.5 | −0.280 | 2.5 (1.6) | 2.8 (1.8) | .017 | 1.4 | 0.177 |
| One-leg stand, s | 10.1 (7.2) | 18.5 (9.3) | <.001 | >1000 | 1.009 | 14.7 (9.3) | 12.1 (8.7) | .015 | 1.9 | −0.290 | 13.1 (9.0) | 15.5 (9.5) | .006 | 3.9 | 0.259 |
| Latent variables | |||||||||||||||
| Locomotion | 2.1 (0.5) | 3.0 (0.5) | <.001 | >1000 | 1.937 | 2.5 (0.6) | 2.3 (0.6) | .001 | 26.8 | −0.317 | 2.4 (0.6) | 2.4 (0.6) | .584 | 0.095 | 0.040 |
| Object manipulation | 5.3 (0.5) | 5.9 (0.5) | <.001 | >1000 | 1.066 | 5.2 (0.5) | 5.0 (0.7) | .003 | 7.2 | −0.277 | 5.2 (0.6) | 4.3 (0.5) | <.001 | >1000 | −1.793 |
| Balance | 1.3 (0.3) | 2.5 (0.7) | <.001 | >1000 | 2.163 | 1.9 (0.8) | 1.5 (0.6) | <.001 | >1000 | −0.522 | 1.7 (0.7) | 2.0 (0.9) | <.001 | >1000 | 0.401 |
Abbreviations: BF, Bayes Factors; BMI, body mass index; SALTO, SALzburg Together against Obesity. Note: Bayesian analyses were conducted using JASP 0.16 (jasp-stats.org) with the default Cauchy prior centered at zero and a scaling parameter of 0.707. Locomotion (range 0–8), Object Manipulation (range 0–6), and Balance (range 1–3) represent the estimated factor scores of the multigroup model.
Intercorrelations of the Latent Variables
| Locomotion and object manipulation | Object manipulation and balance | Locomotion and balance | |||||||
|---|---|---|---|---|---|---|---|---|---|
| r | SE | P | r | SE | P | r | SE | P | |
| Overall | .79 | 0.06 | <.001 | .46 | 0.07 | <.001 | .68 | 0.06 | <.001 |
| Age | |||||||||
| Young | .71 | 0.09 | <.001 | .32 | 0.15 | .032 | .66 | 0.17 | <.001 |
| Old | .75 | 0.09 | <.001 | .36 | 0.12 | .003 | .75 | 0.13 | <.001 |
| BMI | |||||||||
| Normal | .79 | 0.07 | <.001 | .45 | 0.08 | <.001 | .63 | 0.07 | <.001 |
| Overweight | .80 | 0.06 | <.001 | .44 | 0.10 | .010 | .93 | 0.01 | <.001 |
| Sex | |||||||||
| Male | .78 | 0.07 | .080 | .56 | 0.10 | <.001 | .65 | 0.08 | <.001 |
| Female | .84 | 0.08 | <.001 | .48 | 0.10 | <.001 | .70 | 0.09 | <.001 |
Abbreviation: BMI, body mass index; r, Pearson’s correlation coefficient; SE, standard error; P, probability level of significance.
Multigroup Models
All fit indices of the several multigroup models are reported in Table 5. In regard to the age groups, the configural model showed a good fit to the data. Fixing the factor loadings to be equal across age groups (metric invariance) did not lead to a considerable decrease in the model fit indices (ΔCFI = −.009). Hence, the metric invariant model was accepted. In the next step, in a scalar invariant model, the item thresholds and item intercepts were set equal across groups. This led to a decrease in the model fit (ΔCFI = −.033) that was not acceptable according to the chosen criterion. In a partial scalar invariance model, the intercept of the continuous variable Shuttle Run and the threshold of the categorical variable Hopping was estimated freely in both groups, resulting in an acceptable change in the model fit (ΔCFI = −.017). The means and SDs derived from this model are presented in Table 3, and the intercorrelations are in Table 4. Older children showed higher skill levels in all 3 skill domains with large effects (d > 1.066) and decisive evidence (BF10 > 1000) in favor of an age difference (Locomotion: ΔM = 0.91, SE = 0.03; P < .001; Object Manipulation: ΔM = 0.51, SE = 0.04; P < .001; Balance: ΔM = 1.22, SE = 0.04; P < .001). The intercorrelations between Object Manipulation and Locomotion as well as Object Manipulation and Balance did not differ significantly across age groups (Ps > .370, qs < 0.066). The intercorrelation between Locomotion and Balance was stronger in older (r = .75) compared with younger (r = .66) children (P = .014), but with only a small effect size (q = 0.182).
Model Fit Indices of the Multigroup Models by the Levels of Invariance Testing
| Group | Level of invariance | χ2 | df | P | RMSEA | TLI | CFI | ΔCFI |
|---|---|---|---|---|---|---|---|---|
| Age | Configural | 61.4 | 47 | .078 | .029 | .981 | .992 | |
| Metric | 84.1 | 55 | .007 | .038 | .966 | .983 | −.009a | |
| Scalar | 186.8 | 101 | <.001 | .048 | .946 | .950 | −.033b | |
| Scalar partial | 152.2 | 93 | <.001 | .042 | .959 | .966 | −.017b | |
| BMI | Configural | 53.8 | 47 | .231 | .020 | .994 | .997 | |
| Metric | 81.1 | 55 | .013 | .036 | .980 | .990 | −.007a | |
| Scalar | 130.9 | 101 | .024 | .028 | .988 | .989 | −.001b | |
| Sex | Configural | 67.6 | 47 | .026 | .034 | .979 | .991 | |
| Metric | 78.9 | 55 | .019 | .034 | .979 | .990 | −.001a | |
| Scalar | 295.7 | 101 | <.001 | .072 | .908 | .916 | −.074b | |
| Scalar partial | 141.1 | 95 | .002 | .036 | .977 | .980 | −.010b |
Abbreviations: BMI, body mass index; CFI, comparative fit index; RMSEA, root mean square error of approximation; TLI, Tucker–Lewis fit index; ΔCFI, change in CFI; χ2, chi-squared test criterion.
aCompared with the configural model. bCompared with the metric model.
Concerning the BMI groups, the configural model obtained good fit statistics (see Table 5). The restrictions of the metric model did not lead to a considerable decrease in the model fit indices (ΔCFI = −.007). The same was true for testing the scalar invariance (ΔCFI = −.001). Children with overweight were significantly less skilled in all 3 domains compared with children with normal weight. The difference was most pronounced in Balance (ΔM = −0.41, SE = 0.07; P < .001) and Locomotion (ΔM = –0.19, SE = 0.05; P = .001) with a medium (d = −0.522) and a small to medium (d = −0.317) effects size, respectively. The Bayes factors revealed decisive (BF10 > 1000) and strong (BF10 = 26.8) evidence in favor of a group difference. The effect was small (d = −0.277) for Object Manipulation (ΔM = −0.15, SE = 0.05; P = .003) with the data 7.2 times more likely (ie, moderate evidence) under the alternative than the null hypothesis (BF10 = 7.2). The intercorrelations between Object Manipulation and both Locomotion and Balance did not differ significantly across BMI groups (Ps > .605, qs < 0.049). The association between Locomotion and Balance was stronger in children with overweight (r = .93) than in children with normal weight (r = .63; P < .001, q = 0.889).
Finally, concerning sex differences the configural model showed a good fit to the data (see Table 5). Fixing the factor loading to be equal across sex groups (metric invariance) hardly changed the model fit (ΔCFI = −.001). Additionally fixing the item thresholds and item intercepts (scalar invariance) decreased the model fit (ΔCFI = −.074) that was not acceptable according to the chosen criterion. But estimating the threshold of the item Catching freely in both groups kept the model fit change (ΔCFI = −.010) within an acceptable range. Regarding Locomotion skills, no sex differences were found (ΔM = 0.02, SE = 0.04; P = .584). The trivial effect size of d = 0.040 and the Bayes factor of BF10 = 0.095 indicate strong evidence for the absence of a sex effect. Concerning Object Manipulation skills, boys showed significantly higher skill levels than girls (ΔM = −0.97, SE = 0.04; P < .001) with a large effect size (d = −1.793). By contrast, with a medium effect size (d = 0.401), girls turned out to have significantly better Balance skills (ΔM = 0.31, SE = 0.06; P < .001). Bayesian analysis reveals that the factor scores obtained for both Object Manipulation and Balance are over 1000 times more likely under the hypothesis of a difference between boys and girls (BF10 > 1000). The intercorrelations between Object Manipulation and Balance as well as between Locomotion and Balance were not significantly different between boys and girls (Ps > .162, qs < 0.103). With only a small effect size (q = 0.172), Object Manipulation and Locomotion were slightly stronger correlated in girls (r = .84) than in boys (r = .78; P = .020).
Discussion
Our study aimed to address the following questions of FMS in preschoolers: (1) Does the data from the SALTO test battery fit a predefined 3-factor model of Locomotion, Object Manipulation, and Balance? (2) Does the model support a hybrid assessment approach combining process- and product-oriented measures? (3) Is the model equivalent across subgroups of age, weight, and sex? (4) If the model proves to be structurally valid, how do the latent variables differ between younger and older children, between children with normal weight and overweight, and between boys and girls?
The results of the single-indicator multitrait–multimethod analysis confirmed that the construct of FMS in 42- to 72-month-old children has a 3-dimensional structure with Balance as a separate factor in addition to Locomotion and Object Manipulation. The good model fit indices of the confirmatory factor, which included an additional latent method factor, argue for the use of both process- and product-scoring methods to comprehensively capture levels of FMS in preschoolers. The multigroup analyses revealed strong measurement invariance across children with normal weight and overweight, whereas partial strong invariance was found across age and sex groups. Because the noninvariant items constituted only a small portion of the model, we assume they will not significantly bias group comparisons of latent means to any meaningful degree. Therefore, we consider the SALTO model to be reasonably invariant. Group comparisons of the latent constructs confirmed the preschool age as an important period for motor development as older children clearly outperformed younger ones in all skill domains. The results further proved the impact of weight status on the level of FMS as children with overweight had considerably lower skills, particularly in Balance and Locomotion. Sex differences were found in Object manipulation and Balance, with boys showing higher competence when catching, kicking, and throwing is required, whereas girls performed better in the domain of Balance.
Structural Validity of the 3-Factor SALTO Model
That balance acts as a distinct factor of FMS in preschool children becomes evident from the intercorrelations of the latent variables. The correlation between Object Manipulation and Locomotion was found to be larger than their respective interrelationship with Balance. Among these factors, Object Manipulation and Balance had the weakest association, indicating they are the most distinct from each other. This 3-factor model, which includes Locomotion, Object Manipulation, and Balance, aligns with the accepted structure of FMS in children, as advocated from a theoretical viewpoint (27) and supported by empirical evidence (64). In their recent systematic review on observational assessment tools for FMS, Eddy et al (22) identified 9 test batteries that assess all 3 components. Among the 5 instruments aimed at preschool children, 3 exclusively employ a process-oriented and the other 2 a product-oriented approach. The 3 tools that incorporate both static and dynamic balance tasks are those whose structural validity has been evaluated by confirmatory factor analyses, with the number of factors ranging from 2 to 4. The only test battery that has verified a 3-factor model of FMS similar to our SALTO test battery is the Preschooler Gross Motor Quality Scale by Sun et al (71), although it uses only process-oriented measures. More recently, 2 further studies confirmed a 3-dimensional construct of gross motor FMS in preschool children with locomotion, object manipulation, and balance as the latent factors. Aadland et al (1) validated the construct in 2 large samples of Norwegian preschoolers by combining selected locomotion and object manipulation tests from the process-oriented TGMD-3 and 3 process-oriented balance tasks from the Preschooler Gross Motor Quality Scale test battery. Similar to the SALTO assessment tool, Chen et al (15) used both process- and product-oriented measures to assess FMS in Chinese preschool children. The 3-factor structure was evaluated by a principal component analysis, which, however, is ill-suited for construct validation purposes (31). Some authors argue that static and dynamic balance (36), or the ability to cycle a balance bike (39), are task-specific and separable constructs of FMS in early childhood. Our SALTO model, however, supports the concept that balance/stability is not only an underlying ability for many movement skills, but foremost a general and separate subdomain of FMS. Moreover, our study contributes an important aspect to the plethora of FMS assessment tools by confirming that the validity of the 3-factor structure also applies to children with overweight.
Process- and Product-Oriented Testing
Even though our model results do not directly confirm that both process- and product-oriented items are necessarily essential, they support the call for the use of both process- and product-scoring methods to comprehensively capture all levels of FMS in preschoolers. By contrast, the most widely used assessment battery for FMS, the TGMD, which is ascribed the greatest evidence for validity and reliability (22, 81), solely focuses on process-oriented measures and does not include balance-specific tasks. Of the 24 assessment batteries that Eddy et al (22) systematically reviewed, 9 were product-oriented, 13 were process-oriented, and only 2 assessment tools incorporated both process and product methodologies. Similarly, only 10 out of 107 studies included in the review by Hulteen et al (37), which concerned the validity and reliability of FMS assessment tools in children and adolescents, used the hybrid approach. By measuring the form of movements, process-oriented assessments are deemed more suitable for identifying and quantifying developmental delays and for providing a better understanding of the movement strategies employed by children (44). This is particularly relevant for children who are overweight as the movement outcome can be strongly affected by the body weight or physical limitations associated with being overweight. On the other hand, product-oriented measures are more feasible for implementation across various settings as they are less time-consuming, easier to score, and typically require less training (37). These measures are better capable of discriminating advanced skill levels (43) and are not susceptible to bias from preconceptions of what constitutes the ideal movement pattern (14). Because they are less subjective and thus less prone to errors, product-oriented assessments have been shown to provide higher levels of reliability (37). As process- and product-oriented measures assess different aspects of FMS and do not equally evaluate intervention efficacy in preschoolers (55), a hybrid-oriented approach such as the SALTO test battery may better capture multiple salient descriptors of FMS (44). In agreement with Hulteen et al (37), we therefore suggest that assessment tools that combine process- and product-oriented measures yield a more comprehensive evaluation and a holistic understanding of motor competence. However, it remains an open question whether a process-/form- or product-/outcome-oriented approach, or a combination of both, is in fact more sensitive for understanding the relationship between FMS competence and health outcomes and/or for detecting intervention-induced FMS changes.
Measurement Invariance
Another important aspect of test development is the evaluation of measurement invariance as it determines whether a test instrument is appropriate for use in various groups (12). Measurement invariance assesses the equivalence of a construct across groups or measurement occasions and ensures that a construct has the same meaning to those groups (56). Achieving strong measurement invariance is a prerequisite to comparing latent means of FMS across age, sex, and weight status. Consequently, it becomes a fundamental assumption underlying the application of the test batteries in most situations (1). The assumption of strong measurement invariance was met in the SALTO test battery across children with normal weight and overweight, whereas only partial measurement invariance was found across age and sex groups. Younger and older children as well as boys and girls conceptualized the constructs in the same way, and each item contributed to the latent construct to a similar degree across groups, but scalar invariance was only partially supported for age and sex, that is, given the same level of the overall locomotion skills, the item levels of the Shuttle Run (Product) and the Hopping (Process) still differed between younger and older children. At the same time, given the same levels of Object Manipulation skills, the Catching skills differ between boys and girls. Thus, the group differences in the variances but not in the means concerning these variables could be entirely explained by the overall skill factor. However, even in the partial invariant model with the highest number of freely estimated parameters, only 2 parameters had to be estimated without restriction between groups. Compared with the number of 159 parameters in the configural model, which is the model with no restrictions, the number of 2 parameters can be considered as very low, and therefore as unproblematic in terms of biased parameter estimations (16, 66). We therefore assume that the SALTO test battery is suitable for the detection of differences during early childhood development, among BMI groups and between boys and girls in the latent constructs of FMS.
Although measurement invariance is crucial for valid group comparisons, only a few previous studies have tested for it across age or sex. Most recently, Aadland et al (1) employed a multiple-indicators multiple-causes approach with age as the covariate to evaluate measurement invariance between boys and girls and across age in 2 Norwegian samples of preschoolers. They found strong measurement equivalence over age and partial strong invariance between sexes for both a 3-factor model, which included balance, and a simpler 2-factor model that combined locomotion and object manipulation into a single factor. By contrast, by using the same approach, Garn and Webster (25) established strong measurement invariance between 3- to 10-year-old boys and girls but discovered differential item functioning across age for 3 of the 12 TGMD-2 indicators. The same age group was analyzed by Wagner et al (76) for a 2-factor structure of the TGMD-3, which—similar to the TGMD-2—does not include balance-specific tests. They demonstrated only weak invariance across sex and consequently questioned the validity of comparing the latent means between boys and girls, although they did not test for partial scalar invariance. None of these measurement invariance analyses of young children’s FMS accounted for a potential method effect because they were all limited to process-oriented tests. The movement task with an intercept that varied across sex, as analyzed by Aadland et al (1), and across age, as found in our analysis, is Hopping. Therefore, this test may be replaced by an unbiased locomotion task that allows for a more valid comparison of the latent means between groups. Nevertheless, to our knowledge, our validation study is the first to assess the equivalence of FMS constructs across weight status in preschool children, thereby enabling an appropriate and proper comparison of the latent means between children with normal weight and those with overweight.
Group Comparisons
Our results from the comparisons of Locomotion, Object Manipulation, and Balance performance between younger versus older preschoolers, boys versus girls, and children with normal versus increased BMI largely support and further extend previous findings. The largest group effects were observed when comparing the latent means between younger and older kindergarten children, which indicates substantial development in FMS during the kindergarten years. With an effect about twice as large for Locomotion and Balance compared with Object Manipulation, our results are consistent with the outcome of a systematic review on the global levels of FMS in children by Bolger et al (9). This review included data of 64 studies that used the TGMD-2 to assess FMS levels of typically developing children worldwide. It revealed that FMS competence increases across age, ranging from 3 to 10 years, with greater proficiency in locomotor skills than object manipulation skills. The largest increases were detected from ages 3 to 5 in locomotion and 3 to 6 years in object control, indicating that the kindergarten years constitute a highly sensitive period for the maturation of FMS. Our SALTO test battery showed the largest latent mean differences between age groups in the Balance factor, which is not assessed by the TGMD test battery. Our product-oriented static and dynamic balance assessment confirmed the process-oriented findings of Sun et al (71) and Aadland et al (1), who reported moderate to high correlations between chronological age and the quality of motor balance in addition to correlations with locomotion and object manipulation.
Concerning sex differences in FMS, the most consistent finding in the literature is that boys outperform girls on object manipulation skills from childhood (eg, Bolger et al 9) to adolescence (eg, Barnett et al 6), as observed in both process-oriented (eg, Bardid et al 3), and product-oriented (eg, Sember et al 67) test batteries. With the largest sex effect, our results further confirmed that boys perform better than girls in tasks requiring object control. This early sex gap, which is attributed to differing social gender roles, may contribute to girls being less likely to participate in typical male-dominated sports such as soccer or basketball (8). However, further research is needed to understand whether the differences between boys and girls are due to socialization processes or to what extent they may be innate. The discrepancy, however, can be mitigated by customized education programs, as demonstrated by a 6-week intervention study in elementary school children (67). In contrast to the object manipulation factor, previous findings on differences in locomotor proficiency between boys and girls are inconsistent. Some research has reported similar levels of locomotion (eg, 3), other studies have found that girls perform better than boys (eg, 8), and yet others have observed the opposite (eg, 69). In line with Bardid et al (3) and Goodway et al (28), we found no sex-related differences in the Locomotion factor, which—in contrast to previous studies—involved both process-oriented and product-oriented movement tasks. Contrary to Object Manipulation, the SALTO test battery showed that girls outperformed boys on the Balance factor. This result agrees with the findings from Raudsepp and Pääsuke (57), who observed a better outcome-oriented single-leg balance performance in 8-year-old girls compared with age-matched boys. The same observation was also made by Sember et al (67) in 11-year-olds on 2 latent factors of balance, determined by 9 components of 3 static and one dynamic balance tests. The greater balance proficiency among girls may be attributed to the assumption that, from an early age, girls participate more in balance-related types of activities such as dance and gymnastics (cf. 8, 32).
Using a test battery with a verified strong measurement invariance across weight status, the comparisons between BMI groups corroborate previous findings as children with overweight demonstrated lower performance levels in all 3 subdomains of FMS compared with their normal weight counterparts. In their systematic review of 12 cross-sectional studies involving children aged 3–12 years, Slotte et al (68) reported an overall inverse but weak association between FMS and weight status based on 7 studies using process-oriented tests, 4 studies using product-oriented tests, and only one study that used a combination of both assessment approaches. In those studies that employed a 3-dimensional structure of locomotor, object control, and balance skills, locomotion primarily showed a negative relationship with weight status, rather than object control or balance. This point aligns with our results when comparing only Locomotion and Object Manipulation as the latter yielded slightly smaller group differences. It is argued that biomechanical factors related to excessive body weight—such as lower limb problems—may deter children who are overweight or obese from participating in locomotor activities that involve propelling their bodies through space (48). Potential discomfort, additional effort, and self-consciousness about their body image or performance in front of peers may further discourage these children from engaging in sports or activities requiring extensive locomotion. However, unlike Slotte et al.’s review, we observed the largest difference between BMI groups in the Balance factor. This discrepancy could arise from the fact that the 2 indicators for Balance were both product-oriented tests, which are considered to be more sensitive to children’s maturational levels and physical growth than process-oriented assessments. Moreover, all 3 studies that examined the association between balance and weight status were classified by Slotte et al (68) as being of low-quality and hence may have been less likely to detect larger differences. Most recently, a clear inverse association between FMS and BMI was observed by Martins et al (48) in a large sample of 3- to 5-year-old children from 8 countries. In both subdimensions of the TGMD-2, locomotion and object control, the inverse relationship with BMI was even stronger for preschoolers who were classified as obese. Similar results were also found in middle-aged children as those with a higher BMI exhibited lower movement quality in both locomotion/self-movement and object manipulation (40, 46, 77). Furthermore, longitudinal studies revealed that preschool children who changed their weight status from overweight/obese to normal weight displayed, at later ages, FMS comparable to those who remained at normal weight (42), and children with a slower rate of development in their FMS pathways were more prone to become overweight or obese (61). Strong evidence for these bidirectional longitudinal associations between FMS and weight status was also concluded by Barnett et al (7) in their systematic review. Even though BMI-standard deviation score is a generally accepted measure for assessing the weight status of children, it is—as with waist circumference—only a proxy measure and should not be compared with more sophisticated methods of body composition (68). In their recent review, Liu et al (43) differentiated between these measures and reported consistent negative correlations between body composition and FMS in children and adolescents in studies using more accurate measures of total body or abdominal fat, whereas the use of BMI and waist circumference as evaluation criteria yielded inconsistent results. We further argue that test instruments for FMS—such as the SALTO test battery—should be complemented by more advanced methodologies such as nonlinear tools (eg, sample entropy, largest Lyapunov exponent, or detrended fluctuation analysis) to unfold regulatory mechanism of movement coordination. This could prove particularly useful in the context of postural balance and weight status, as recently shown by Wiesinger et al (78). Using such tools, they observed higher postural fluctuations and reduced structural complexity indicating impaired regulatory mechanisms in children and adolescents with overweight or obesity compared with peers with normal weight. The routine deployment of such methods is now facilitated by the ongoing advancements in technology through the use of smartphones and inertial measurement units. Nevertheless, the outcome of our BMI group comparisons aligns with previous findings that suggest a possible role of FMS in maintaining and achieving a healthy weight throughout childhood, presumably mediated through engagement in physical activity.
Limitations
The SALTO test battery did not allow for the employment of a multiple-indicator model on the data set because multiple-indicator approaches require at least 2 indicators for each trait-method unit. This assumption was only met by Locomotion but not by Object Manipulation (assessed only process-oriented) and Balance (assessed only product-oriented). Single-indicator models have been criticized as they might lead to biased estimates of convergent validity, method specificity, and reliability, especially if the actual correlations of method factors across traits are low (23, 26). However, across research fields, the applications of multiple-indicator models are still rare (26) and, in the context of efficiently testing movement skills, practically difficult or even impossible to implement, not least because of the limited attention span in young children. Nevertheless, to our knowledge, we are the first to demonstrate measurement equivalence in an FMS test battery across essential subgroups of kindergarten children based on a multitrait–multimethod model. This is an important step forward for valid comparisons and interpretations of FMS between sex, age, and BMI groups. It should also be noted that in the SALTO test battery only the Broad Jump test follows a true hybrid approach, in which the same test item has both process and product criteria for each task. All other tests involve measures that are either product-oriented or process-oriented. The selection of the respective criteria was based on the rationale of which measure better assesses the children’s skill performance without disadvantaging children due to their overweight status while minimizing the time and/or personnel effort required, thus enabling the practical implementation of the test battery in applied settings. However, further research should address this limitation. Another limitation of the applied multigroup confirmatory factor analysis is that it did not account for potential interaction effects between group variables. In order to weaken this limitation, we further evaluated whether the subgroup characteristics were equal across each other. Although the children’s age did not differ between boys and girls (P = .933; BF10 = 0.083; d < 0.01) and the standardized BMI was not different between younger and older children (P = .133; BF10 = 0.249; d = 0.11), boys were found to have higher BMI-standard deviation score values than girls (P = .006; BF10 = 3.6; d = 0.21). However, because the effect is small, we assume that interaction effects play—if at all—only a minor role.
Conclusion
The SALTO test battery fulfils the current call that valid assessment tools for FMS should encompass locomotor skills, object control, and balance and should follow a hybrid approach of combining process-oriented and product-oriented measures. Our confirmatory factor analyses and measurement invariance testing demonstrated that the SALTO test battery is a structurally valid tool for evaluating and comparing FMS between BMI weight status categories among male versus female kindergarten children from 3 to 6 years. Against the background of worldwide increasing obesity among children and adults, the main value of our study is to confirm that the 3-factor structure of the SALTO model also holds for preschool children who are overweight. A further strength of our study is its methodological quality, which is classified as very good according to the COSMIN Risk of Bias checklist (48). Our results of the group comparisons underscore the importance of considering age, body weight, and sex differences when designing interventions or educational programs aimed at enhancing FMS in early childhood. We reasonably suggest that the SALTO test battery is well suited to assessing the relationships between FMS, weight status, development, and health parameters. As FMS are considered to build the foundation of more advanced, complex movements required to participate in context-specific movement activities, they are reciprocally related to physical activity (30, 70). Physical activity, in turn, is considered to be an important strategy in preventing and decreasing childhood obesity (20, 38). Therefore, an early emphasis on promoting FMS may help to break the negative spiral of low motor skill competence, reduced physical activity, and increased weight and obesity (71). Similarly, encouraging the early improvement of object manipulation skills in girls and balance skills in boys may help close—or at least narrow—the sex gap that otherwise persists into adolescence (6, 67). Moreover, implementing interventions to enhance FMS in early childhood through the use of targeted testing procedures, such as the SALTO test battery, may also contribute to physical, psychological, social, and academic (health) benefits because they have been found to be positively associated with motor skill competence (35). These and our findings have important implications for educators, parents, and health care professionals working with preschool children as they highlight the need for tailored approaches in promoting physical development and addressing disparities in movement skill proficiency among young learners. Professionals who work with children should use valid and feasible test batteries to identify specific weaknesses and strengths in the child’s FMS, which in turn can guide personalized interventions, ensuring that each child receives the appropriate level of support and challenge. This individualized approach better supports children in their motor learning processes and maximizes the potential for improvement and engagement. In addition, regular assessments allow professionals to track progress over time and adjust programs as needed, helping to maintain the relevance and effectiveness of the interventions. Future research should focus on different types of reliability as well as concurrent, cross-cultural, and predictive validity to further enhance the utility and validity of the SALTO assessment tool for the evaluation of FMS in pediatric populations.
Acknowledgment
The authors thank Johannes Dirnberger for proofreading and constructive contribution to the manuscript.
References
- 1.↑
Aadland KN, Nilsen AKO, Lervag AO, Aadland E. Structural validity of a test battery for assessment of fundamental movement skills in Norwegian 3–6-year-old children. J Sports Sci. 2022;40(15):1688–99. doi:
- 2.↑
Arbuckle JL. Full information estimation in the presence of incomplete data. In: Marcoulides GA, Schumacker RE, editors. Advanced Structural Equation Modeling. 1st ed. Psychology Press; 1996. pp. 243–77.
- 3.↑
Bardid F, Huyben F, Lenoir M, et al. Assessing fundamental motor skills in Belgian children aged 3–8 years highlights differences to US reference sample. Acta Paediatr. 2016;105(6):e281–90. doi:
- 4.↑
Barnett LM, Lai SK, Veldman SLC, et al. Correlates of gross motor competence in children and adolescents: a systematic review and meta-analysis. Sports Med. 2016;46(11):1663–1688. doi:
- 5.↑
Barnett LM, Stodden D, Cohen KE, et al. Fundamental movement skills: an important focus. J Teach Phys Educ. 2016;35(3):219–225. doi:
- 6.↑
Barnett LM, van Beurden E, Morgan PJ, Brooks LO, Beard JR. Gender differences in motor skill proficiency from childhood to adolescence: a longitudinal study. Res Q Exerc Sport. 2010;81(2):162–70.
- 7.↑
Barnett LM, Webster EK, Hulteen RM, et al. Through the looking glass: a systematic review of longitudinal evidence, providing new insight for motor competence and health. Sports Med. 2022;52(4):875–920. doi:
- 8.↑
Bolger L, Bolger L, O’Neill C, et al. Age and sex differences in fundamental movement skills among a cohort of Irish school children. J Mot Learn Dev. 2018;6(1):81–100. doi:
- 9.↑
Bolger L, Bolger L, O’Neill C, et al. Global levels of fundamental motor skills in children: a systematic review. J Sports Sci. 2021;39(7):717–753. doi:
- 11.↑
Bös K, Bappert S, Tittlbach S, et al. Karlsruher Motorik-Screening für Kindergartenkinder (KMS 3-6). Sportunterricht. 2004;53(3):79–87.
- 12.↑
Brown TA. Confirmatory Factor Analysis for Applied Research. 2nd ed. Guilford Publications; 2015.
- 14.↑
Capio CM, Sit CH, Abernethy B. Fundamental movement skills testing in children with cerebral palsy. Disabil Rehabil. 2011;33(25–26):2519–28. doi:
- 15.↑
Chen Y, Gu Y, Tian Y, et al. Developing a scale for measuring the fundamental movement skills of preschool children in China. Int J Environ Res Public Health. 2022;19(21):14257. doi:
- 16.↑
Cheung GW, Rensvold RB. Evaluating goodness-of-fit indexes for testing measurement invariance. Struct Equ Modeling. 2002;9(2):233–55. doi:
- 17.↑
Chew-Bullock TS, Anderson DI, Hamel KA, Gorelick ML, Wallace SA, Sidaway B. Kicking performance in relation to balance ability over the support leg. Hum Mov Sci. 2012;31(6):1615–23. doi:
- 18.↑
Clark JE. On the problem of motor skill development. J Phys Educ Recreat Dance. 2007;78(5):39–44. doi:
- 19.↑
Deoni SC, Dean DC, Remer J, Dirks H, O’Muircheartaigh J. Cortical maturation and myelination in healthy toddlers and young children. Neuroimage. 2015;115(15):147–61. doi:
- 20.↑
Di Maglie A, Marsigliante S, My G, Colazzo S, Muscella A. Effects of a physical activity intervention on schoolchildren fitness. Physiol Rep. 2022;10(2):e15115. doi:
- 21.↑
Donnelly FC, Mueller SS, Gallahue DL. Developmental Physical Education for All Children: Theory Into Practice. Human Kinetics; 2017.
- 22.↑
Eddy LH, Bingham DD, Crossley KL, et al. The validity and reliability of observational assessment tools available to measure fundamental movement skills in school-age children: a systematic review. PLoS One. 2020;15(8):e0237919. doi:
- 23.↑
Eid M, Lischetzke T, Nussbeck FW, Trierweiler LI. Separating trait effects from trait-specific method effects in multitrait-multimethod models: a multiple-indicator CT-C(M-1) model. Psychol Methods. 2003;8(1):38–60. doi:
- 24.↑
Gallahue DL, Ozmun JC, Goodway J. Understanding Motor Development: Infants, Children, Adolescents and Adults. 7th ed. McGraw-Hill; 2012.
- 25.↑
Garn AC, Webster EK. Reexamining the factor structure of the test of gross motor development—second edition: application of exploratory structural equation modeling. Meas Phys Educ Exerc. 2018;22(3):200–12. doi:
- 26.↑
Geiser C, Simmons TG. Do method effects generalize across traits (and what if they don’t)? J Pers. 2021;89(3):382–401. doi:
- 27.↑
Goodway JD, Ozmun JC, Gallahue DL. Understanding Motor Development: Infants, Children, Adolescents, Adults. Jones & Bartlett Learning; 2019.
- 28.↑
Goodway JD, Robinson LE, Crowe H. Gender differences in fundamental motor skill development in disadvantaged preschoolers from two geographical regions. Res Q Exerc Sport. 2010;81(1):17–24. doi:
- 29.↑
Graham JW. Missing data analysis: making it work in the real world. Annu Rev Psychol. 2009;60(1):549–76. doi:
- 30.↑
Graham M, Azevedo L, Wright M, Innerd AL. The effectiveness of fundamental movement skill interventions on moderate to vigorous physical activity levels in 5- to 11-year-old children: a systematic review and meta-analysis. Sports Med. 2022;52(5):1067–90. doi:
- 31.↑
Gruijters SLK. Using principal component analysis to validate psychological scales: bad statistical habits we should have broken yesterday. Eur J Health Psychol. 2019;20(5):544–9.
- 32.↑
Hardy LL, King L, Farrell L, Macniven R, Howlett S. Fundamental movement skills among Australian preschool children. J Sci Med Sport. 2010;13(5):503–8. doi:
- 34.↑
Herrmann C, Heim C, Seelig H. Diagnosis and development of basic motor competencies. Z Entwicklungspsychol Padagog Psychol. 2017;49(4):173–85.
- 35.↑
Hill PJ, McNarry MA, Mackintosh KA, et al. The influence of motor competence on broader aspects of health: a systematic review of the longitudinal associations between motor competence and cognitive and social-emotional outcomes. Sports Med. 2024;54(2):375–427. doi:
- 36.↑
Hirata S, Kita Y, Yasunaga M, et al. Applicability of the movement assessment battery for children-second edition (MABC-2) for Japanese children aged 3–6 years: a preliminary investigation emphasizing internal consistency and factorial validity. Front Psychol. 2018;9:1452. doi:
- 37.↑
Hulteen RM, Barnett LM, True L, Lander NJ, Del Pozo Cruz B, Lonsdale C. Validity and reliability evidence for motor competence assessments in children and adolescents: a systematic review. J Sports Sci. 2020;38(15):1717–98. doi:
- 38.↑
Jurado-Castro JM, Gil-Campos M, Gonzalez-Gonzalez H, Llorente-Cantarero FJ. Evaluation of physical activity and lifestyle interventions focused on school children with obesity using accelerometry: a systematic review and meta-analysis. Int J Environ Res Public Health. 2020;17(17):6031. doi:
- 39.↑
Kavanagh JA, Issartel J, Moran K. Quantifying cycling as a foundational movement skill in early childhood. J Sci Med Sport. 2020;23(2):171–5. doi:
- 40.↑
Kelly L, O’Connor S, Harrison AJ, Ní Chéilleachair NJ. Does fundamental movement skill proficiency vary by sex, class group or weight status? Evidence from an Irish primary school setting. J Sports Sci. 2019;37(9):1055–63. doi:
- 41.↑
Krombholz H. Testbatterie zur Erfassung motorischer Leistungen im Vorschulalter MoTB 3–7. Beschreibung, Gütekriterien, Normwerte und ausgewählte Ergebnisse. 2011. http://hdl.handle.net/20.500.11780/3358
- 42.↑
Lima RA, Soares FC, Queiroz DR, Aguilar JA, Bezerra J, Barros MVG. The importance of body weight status on motor competence development: from preschool to middle childhood. Scand J Med Sci Sports. 2021;S1 31:15–22. doi:
- 43.↑
Liu C, Cao Y, Zhang Z, Gao R, Qu G. Correlation of fundamental movement skills with health-related fitness elements in children and adolescents: a systematic review. Front Public Health. 2023;11:1129258. doi:
- 44.↑
Logan SW, Barnett LM, Goodway JD, Stodden DF. Comparison of performance on process- and product-oriented assessments of fundamental motor skills across childhood. J Sports Sci. 2017;35(7):634–41. doi:
- 45.↑
Longmuir PE, Boyer C, Lloyd M, et al. Canadian Agility and Movement Skill Assessment (CAMSA): validity, objectivity, and reliability evidence for children 8–12 years of age. J Sport Health Sci. 2017;6(2):231–40. doi:
- 46.↑
Maïano C, Morin AJS, April J, et al. Psychometric properties of a French-Canadian version of the test of gross motor development—third edition (TGMD-3): a bifactor structural equation modeling approach. Meas Phys Educ Exerc. 2022;26(1):51–62. doi:
- 47.↑
Marsh HW, Byrne BM, Craven R. Overcoming problems in confirmatory factor analyses of MTMM data: the correlated uniqueness model and factorial invariance. Multivariate Behav Res. 1992;27(4):489–507. doi:
- 48.↑
Martins C, Romo-Perez V, Webster EK, et al. Motor competence and body mass index in the preschool years: a pooled cross-sectional analysis of 5545 children from eight countries. Sports Med. 2024;54(2):505–16. doi:
- 49.
Mokkink LB, de Vet HCW, Prinsen CAC, et al. COSMIN risk of bias checklist for systematic reviews of patient‐reported outcome measures. Qual Life Res. 2017;27(5):1171–9. doi:
- 50.↑
Moller S, Poulain T, Korner A, et al. Motor skills in relation to body-mass index, physical activity, TV-watching, and socioeconomic status in German four-to-17-year-old children. PLoS One. 2021;16(5):e0251738.
- 51.↑
Morrow JR, Zhu W, Franks BD, Meredith MD, Spain C. 1958–2008: 50 years of youth fitness tests in the United States. Res Q Exerc Sport. 2009;80(1):1–11.
- 53.↑
Nardini M, Cowie D. The development of multisensory balance, locomotion, orientation, and navigation. In: Bremner AJ, Lewkowicz DJ, Spence C, editors. Multisensory Development. Oxford University Press; 2012. pp. 137–58.
- 54.↑
Ng JL, Button C. Reconsidering the fundamental movement skills construct: implications for assessment. Mov Sport Sci. 2019;(102):19–29.
- 55.↑
Palmer KK, Stodden DF, Ulrich DA, Robinson LE. Using process- and product-oriented measures to evaluate changes in motor skills across an intervention. Meas Phys Educ Exerc Sci. 2021;25(3):273–82. doi:
- 56.↑
Putnick DL, Bornstein MH. Measurement invariance conventions and reporting: the state of the art and future directions for psychological research. Dev Rev. 2016;41:71–90. doi:
- 57.↑
Raudsepp L, Pääsuke M. Gender differences in fundamental movement patterns, motor performances, and strength measurements of prepubertal children. Pediatr Exerc Sci. 1995;7(3):294–304. doi:
- 58.↑
Remer J, Croteau-Chonka E, Dean DC, et al. Quantifying cortical development in typically developing toddlers and young children, 1–6 years of age. Neuroimage. 2017;153:246–61. doi:
- 59.↑
Ring-Dimitriou S, Freudenthaler T, Aistleitner V, et al. SALTO—Study protocol and rationale of a community-oriented obesity prevention program in the kindergarten. Obes Facts. 2018;11(3):234–46. doi:
- 60.↑
Robinson LE, Stodden DF, Barnett LM, et al. Motor competence and its effect on positive developmental trajectories of health. Sports Med. 2015;45(9):1273–84. doi:
- 61.↑
Rodrigues LP, Stodden DF, Lopes VP. Developmental pathways of change in fitness and motor competence are related to overweight and obesity status at the end of primary school. J Sci Med Sport. 2016;19(1):87–92. doi:
- 62.↑
Rosnow RL. Effect sizes for experimenting psychologists. Can J Exp Psychol. 2003;57(3):221–37. doi:
- 63.↑
Rudd JR, Barnett LM, Butson ML, Farrow D, Berry J, Polman RC. Fundamental movement skills are more than run, throw and catch: the role of stability skills. PLoS One. 2015;10(10):e0140224. doi:
- 64.↑
Rutkowski L, Svetina D. Assessing the hypothesis of measurement invariance in the context of large-scale international surveys. Educ Psychol Meas. 2014;74(1):31–57. doi:
- 65.↑
Schafer JL, Graham JW. Missing data: our view of the state of the art. Psychol Methods. 2002;7(2):147–77. doi:
- 66.↑
Schmitt N, Golubovich J, Leong FT. Impact of measurement invariance on construct correlations, mean differences, and relations with external correlates: an illustrative example using Big Five and RIASEC measures. Assessment. 2011;18(4):412–27. doi:
- 67.↑
Sember V, Groselj J, Pajek M. Balance tests in pre-adolescent children: Retest reliability, construct validity, and relative ability. Int J Environ Res Public Health. 2020;17(15):5474–87. doi:
- 68.↑
Slotte S, Sääkslahti A, Kukkonen-Harjula K, Rintala P. Fundamental movement skills and weight status in children: a systematic review. Balt J Health Phys Act. 2017;9(2):115–27. doi:
- 69.↑
Spessato B, Gabbard C, Valentini N, Rudisill M. Gender differences in Brazilian children’s fundamental movement skill development. Early Child Dev Care. 2013;183(7):916–23. doi:
- 70.↑
Stodden DF, Goodway JD, Langendorfer SJ, et al. A developmental perspective on the role of motor skill competence in physical activity: an emergent relationship. Quest. 2008;60(2):290–306. doi:
- 71.↑
Sun SH, Zhu YC, Shih CL, Lin CH, Wu SK. Development and initial validation of the preschooler gross motor quality scale. Res Dev Disabil. 2010;31(6):1187–96. doi:
- 72.↑
Tittlbach S, Bappert S, Bös K, Woll A. Karlsruher Motorik-Screening für Kindergartenkinder (KMS 3–6). Sportunterricht. 2004;53(3):79–87.
- 73.↑
Tyler R, Foweather L, Mackintosh KA, Stratton G. A dynamic assessment of children’s physical competence: the Dragon Challenge. Med Sci Sports Exerc. 2018;50(12):2474–87. doi:
- 74.↑
Ulrich DA. Introduction to the special section: evaluation of the psychometric properties of the TGMD-3. J Mot Learn Dev. 2017;5(1):1–4. doi:
- 75.↑
Ulrich DA. TGMD-2 Test of Gross Motor Development Second Edition Examiners Manual. Pro-Ed; 2000.
- 76.↑
Wagner MO, Webster EK, Ulrich DA. Psychometric properties of the test of gross motor development 3rd edition (German translation)—results of a pilot-study. J Mot Learn Dev. 2017;5(1):29–44. doi:
- 77.↑
Wälti M, Seelig H, Adamakis M, et al. Investigating levels and determinants of primary school children’s basic motor competencies in nine European countries. Zeitschrift für Grundschulforschung. 2023;16(1):113–33. doi:
- 78.↑
Wiesinger HP, Buchecker M, Müller E, Stöggl T, Birklbauer J. Decreased postural complexity in overweight to obese children and adolescents: a cross-sectional study. Front Hum Neurosci. 2022;16:850548. doi:
- 79.↑
Williams HG, Pfeiffer KA, O’Neill JR, et al. Motor skill performance and physical activity in preschool children. Obesity. 2008;16(6):1421–6. doi:
- 80.↑
World Health Organization. WHO Child Growth Standards. Methods and Development. Length/Height-For-Age, Weight-For-Age, Weight-For-Length, Weight-For-Height and Body Mass Index-For-Age. WHO; 2006.
- 81.↑
World Health Organization. Obesity and overweight. 2021 [cited October 25, 2021]. Available from https://www.who.int/en/news-room/fact-sheets/detail/obesity-and-overweight
- 82.↑
Xin F, Chen ST, Clark C, Hong JT, Liu Y, Cai YJ. Relationship between fundamental movement skills and physical activity in preschool-aged children: a systematic review. Int J Environ Res Public Health. 2020;17(10):3566. doi:
- 83.
Zimmer R, Volkamer M. MOT 4–6. Motoriktest für vier- bis sechsjährige Kinder. Beltz Test Gesellschaft; 1984.
