Preliminary evidence supports the validity of the Strength-Dexterity Test as a unidimensional scale measuring fingertip force coordination in children and adolescents

Research objective: To evaluate internal scale validity, concurrent validity and age and gender effects of the Strength‐Dexterity Test (S‐DT) in children/adolescents. Design: Observational study of measurement properties of the S‐DT using Rasch measurement modelling and hypothesis testing. Participants: Fifty‐six typically developing children/adolescents aged 4–17 years (mean 9 years 8 months), 30 males and 26 females, 54 were right‐handed. Setting: Sweden. Measures: S‐DT aims to measure coordinated finger‐tip forces and consists of 82 compression springs requiring varying pinch forces to compress (strength) the spring shape without distortion (dexterity). Subsets of 53–82 springs, requiring strength just above and below each participant’s maximal pinch force, were presented in random order for compression using the dominant hand. Success (score 1) was recorded if at least one of three trials achieved full, undistorted compression using correct finger posture. Testing typically took 15 minutes. Additional measures included the Box and Blocks Test of Manual Dexterity (B&B) and a pinch gauge testing maximal pinch strength. Procedures: Tests were presented in the following order: B and B, pinch strength and S‐DT. Rasch measurement modelling for dichotomous data (Winsteps 3.65.0) assessed: (i) internal scale validity including unidimensionality of items; (ii) spread of items and people along the continuum; and (iii) sensitivity of items to differences among participants. Reliability was assessed by testing item ability to separate participants into distinct levels of ability. Regression analyses assessed shared variance amongst the three measures and ANOVA tested for gender and age effects. Main findings: Four S‐DT items not mastered by any participant were removed after initial analysis. Twenty items achieved by all participants were retained as the test is ultimately required for those with impairments. The remaining 78 items demonstrated good fit to the model (