The Specificity of Motor Ability Hypothesis states that motor skills are specific to a particular task and are relatively independent from each other. These small to moderate correlations can be interpreted as evidence in favor of the Specificity of Motor Abilities Hypothesis, since a General Motor Ability would have resulted in high correlations for all tasks. Our results are in line with conclusions from other authors that motor skills have a moderate degree of generalizability and are partially task-specific.

In general, low pair-wise correlations (r² < 0.20) between the different motor tasks were found. The highest correlation was between the placing bricks and building bricks r = 0.45 (TMC); the stork balance and jumping in squares r = 0.45 (MABC). The smallest correlation was found between MABC flower trail and throwing bean bag into box (r = 0.02) and between TMC placing bricks and walking/running in slopes (r = 0.02). These low pair-wise relations of items are consistent with findings from younger and older children's age-related motor competence test batteries and provide evidence with children for specificity rather than generality in learning motor skills.

The concept of a general motor ability factor has been challenged by empirical work showing that correlations among different motor tasks are often low. In Fleishman’s and later meta-analytic work, intercorrelations between specific motor abilities and across diverse motor tasks typically fell in the low-to-moderate range, leading many authors to conclude that performance in one motor task only weakly predicts performance in other, dissimilar motor tasks. The paper reviews studies in which batteries of motor tasks show only limited shared variance, arguing against a strong, unitary motor ability factor and limiting the predictive power of single-task measures for overall motor performance.

Horgan and Morgan measured performance in chess and in a complex manual tracking task and found essentially no relationship between the two. They report that "correlations between chess skill and performance on the pursuit rotor were near zero," concluding that skill in one complex domain did not generalize to the other. The authors argue that this dissociation illustrates the task-specific nature of both cognitive and motor skills and the limited predictive value of one task for another.

The authors report that, "High correlations were found for test-retest correlations of the same motor task but only low correlations across different motor tasks despite identical sample and time of measurement." They conclude, "Our results suggest that individuals have different abilities across motor tasks...Consequently, overall motor competence may not be indicated adequately by a single (or few) motor tasks." This directly supports the idea that inter-task correlations are low, limiting the predictive validity of single-task assessments.

Seashore (1930), for example, tested 50 adults on eight fine motor skills and found only weak correlation values (averaging 0.25) between them. Henry (1961) compared two hypothesised specific motor abilities, ‘reaction time’ and ‘speed of movement’ and found almost zero correlation between them. Other studies comparing tasks of balance (Bachman, 1961; Drowatzky & Zuccato, 1967) and strength (Berger, 1962) found similarly low correlations. To date, first order factor analyses of motor performance data derived from multiple tasks have seemed to provide the principal support for the existence of multiple motor abilities, distinct from a global motor ability, since, as noted above, only weak correlations between separate motor skills are typically found.

The abstract states that the study "tested these contentions by assessing the extent of relationship between fine motor tasks, using correlations between selected performance measures across the three tasks." It notes that "intertask correlations were generally low to moderate," indicating that performance on one fine motor task did not strongly predict performance on other fine motor tasks in young adults.

Using a broad test battery in 3- to 5-year-olds, Hestbaek et al. reported that "the correlations between the different motor performance tests were mostly low to moderate." They found that a multi-factor model with separate factors for balance, object control, and locomotion fit better than a single-factor model, leading them to conclude that "motor ability in preschool children is multidimensional" and that no single test captured overall motor performance adequately.

The study examined how well an early single motor assessment predicts later overall motor proficiency. Controlling for age, the correlation between TIMP scores in infancy and BOTMP scores at school age was r = 0.36, described as a moderate relationship. The authors report that TIMP scores, perinatal risk (POPRAS), and age together accounted for 37% of the variance in later BOTMP motor scores, indicating that a single early motor test explains only a limited portion of overall later motor performance.

In a set of interception and timing tasks, the authors report that "inter-task correlations revealed that the tasks tested different aspects of predictive processes." They emphasize that despite all being predictive motor tasks, performance did not strongly covary: "the pattern of results suggests only weak to moderate correlations between tasks," supporting the view that predictive motor ability is not a single, unitary construct across tasks.

This systematic review reports that studies using batteries of motor tests frequently find weak to moderate correlations among different motor tasks, suggesting a multidimensional rather than unitary construct of motor competence. The authors highlight that because distinct tasks load on different factors, prediction of global motor competence from any single test is limited, and comprehensive assessment requires multiple tasks representing different domains.

The hypothesis that tasks are specific and not based upon general underlying factors (specificity hypothesis) was tested for motor response consistency, or intra-individual variability, SD(intra), and inter-correlations among tasks. Response consistency scores were very low indicating specificity of response consistency. The findings generally support the concept of specificity in the learning and performance of skills; however, low correlations were found between disc-tossing performance and proficiency in the skills of hopscotch and mirror-tracing.

Using multiple motor tasks, the authors tested for a single general motor factor. They found that correlations between different motor tasks were mostly weak to moderate and that a one-factor model fit the data poorly compared with multifactor models. They note that this pattern limits the utility of any single task as a robust predictor of broad motor performance, particularly in heterogeneous populations.

The study reports that "correlations between different field tests of agility, sprint speed and coordination were only low to moderate," despite all being lower-limb, sport-related motor tasks. The authors state that this pattern "supports the view that these qualities are relatively independent," and caution that using a single field test provides only a limited indication of a player's overall motor performance profile.

This paper examined correlations among a battery of motor tasks and reported that "intertask correlations tended to be low, with few coefficients exceeding 0.40." The authors note that such a correlation pattern "is inconsistent with a strong general motor ability factor" and implies that "performance on a single motor test has limited value for predicting performance on other, dissimilar motor tasks."

Summarizing validation work on the AMPS, the database notes that the AMPS Motor score shows only "poor ability ... to predict ADL motor performance (AUC = .68)" in one large psychiatric sample, whereas prediction of other outcomes is better. It also states that correlations between motor and process scores and different aspects of daily functioning range from adequate to excellent, illustrating that "a single motor score" only partially predicts performance across the range of everyday motor tasks assessed.

Across the classic motor behavior literature, multiple studies (e.g., Seashore’s fine motor tests, Henry’s reaction time vs. movement speed, and Drowatzky & Zuccato’s balance tasks) consistently report average correlations in the range of about 0.00–0.30 between different motor tasks. These weak correlations led many authors to argue that a score on any single motor task has limited predictive validity for a person’s overall motor performance profile, motivating the task-specificity view of motor abilities.

The general motor ability hypothesis is that abilities are highly related and can be grouped as a single overall ability. The specificity hypothesis is that abilities are independent and a person's profile consists of amounts of each specific ability. Research evidence supports this hypothesis, as studies have found abilities like reaction time and movement speed to be uncorrelated. Henry and colleagues (1960) showed low correlations between RT and MT, which indicates that performance on one motor task does not strongly predict performance on another.