A Non-Parametric Test-Retest Analysis of Items from a Novel Self-Report Balance Confidence Inventory
Presentation Type
Oral Presentation
Presentation Type
Submission
Keywords
psychometric, reliability, scale, stability, postural control, self-efficacy, self-perception
Department
Sports Medicine
Major
Sports Medicine
Abstract
Balance confidence is a proven factor of health in geriatric and neurological populations (e.g., fall risk, functional mobility, quality of life). However, in healthy, younger-to-middle-aged adults, the measurement (and subsequently the understanding) of balance confidence is scant. The purpose of this study was to investigate the test-retest properties of 14 individual items from a novel, self-report balance confidence inventory. Using various recruitment techniques, a sample of non-pathological adults completed an electronic Qualtrics survey which included the newly developed balance confidence inventory (N=1,259; 18-62 years old, Mage=22.8, SD=5.5; 70% women). Thereafter, 27% of the sample completed the inventory a second time (i.e., retest; N=345; 18-52 years old, Mage=21.3, SD=4.4; 74% women). Participants were required to wait at least seven (but no more than 31) days to complete the inventory a second time. To examine the test-retest reliability of the 14 individual balance confidence inventory items, Svensson’s non-parametric methods for paired ordinal data were used. Per Svensson’s methodology, the calculated statistics were percent agreement, relative position, relative concentration, relative rank invariance, internal rank variance, augmented rank-order agreement coefficient, and coefficient of monotonic agreement. All analyses were completed in Excel and/or RStudio. Thirteen of the 14 items presented with consistent and stout test-retest properties. One item (i.e., parkour/freerunning) presented with mixed (but adequate) test-retest properties. Potential and future adopters should have strong confidence in the item-level test-retest properties of the newly introduced self-report balance confidence inventory. Additional research in line with the current investigation is advised (e.g., replication, different populations, extended test-retest intervals).
Faculty Mentor
Dr. Adam Pennell
Location
Black Family Plaza Classroom 190
Start Date
11-4-2025 2:15 PM
End Date
11-4-2025 2:30 PM
A Non-Parametric Test-Retest Analysis of Items from a Novel Self-Report Balance Confidence Inventory
Black Family Plaza Classroom 190
Balance confidence is a proven factor of health in geriatric and neurological populations (e.g., fall risk, functional mobility, quality of life). However, in healthy, younger-to-middle-aged adults, the measurement (and subsequently the understanding) of balance confidence is scant. The purpose of this study was to investigate the test-retest properties of 14 individual items from a novel, self-report balance confidence inventory. Using various recruitment techniques, a sample of non-pathological adults completed an electronic Qualtrics survey which included the newly developed balance confidence inventory (N=1,259; 18-62 years old, Mage=22.8, SD=5.5; 70% women). Thereafter, 27% of the sample completed the inventory a second time (i.e., retest; N=345; 18-52 years old, Mage=21.3, SD=4.4; 74% women). Participants were required to wait at least seven (but no more than 31) days to complete the inventory a second time. To examine the test-retest reliability of the 14 individual balance confidence inventory items, Svensson’s non-parametric methods for paired ordinal data were used. Per Svensson’s methodology, the calculated statistics were percent agreement, relative position, relative concentration, relative rank invariance, internal rank variance, augmented rank-order agreement coefficient, and coefficient of monotonic agreement. All analyses were completed in Excel and/or RStudio. Thirteen of the 14 items presented with consistent and stout test-retest properties. One item (i.e., parkour/freerunning) presented with mixed (but adequate) test-retest properties. Potential and future adopters should have strong confidence in the item-level test-retest properties of the newly introduced self-report balance confidence inventory. Additional research in line with the current investigation is advised (e.g., replication, different populations, extended test-retest intervals).
