Real-Time Versus Video-Recorded Step Count Validation Among Young Adults

Area of Science:

• Exercise Physiology
• Biomechanical Analysis
• Wearable Technology Validation

Background:

• Consumer Technology Association guidelines suggest dual researcher review of treadmill step counts from video.
• Current post hoc validation methods are resource-intensive, raising questions about necessity.
• This study investigates the accuracy and efficiency of direct observation versus video review for step counting.

Purpose of the Study:

• To evaluate the accuracy of step counts obtained through direct observation.
• To assess the efficiency of direct observation compared to video recording reviews.
• To determine if current video review guidelines are necessary for treadmill step counting.

Main Methods:

• 72 participants (ages 18-20) completed treadmill bouts at various speeds (0.2-2.7 m/s).
• Steps were directly observed and hand-tallied in real-time.
• A subset of video recordings underwent two independent recounts, yielding three total counts per bout.

Main Results:

• Excellent interrater reliability (ICC ≥ .995) was observed across all speeds.
• The average percent deviation between counts was minimal (0.29%), with bout-specific deviations under 1%.
• High correlation between absolute count differences and average counts was found.

Conclusions:

• Single, real-time direct step counts are sufficient and efficient for controlled treadmill settings.
• Video recounts can serve as a backup for data integrity and random verification.
• Findings support optimizing personnel time and reducing costs in large-scale treadmill studies.