Absolute error

Area of Science:

• Motor Control and Learning
• Quantitative Psychology
• Educational Measurement

Background:

• Performance analysis in motor learning often relies on error metrics.
• Existing measures like algebraic error (CE), absolute error (AE), and within-subject variance (VE) lack unambiguous statistical definitions.
• Misinterpretation of these variables can lead to flawed conclusions in research.

Purpose of the Study:

• To re-examine and clarify the statistical and logical meanings of CE, AE, and VE.
• To establish unambiguous definitions for these key performance variables.
• To provide a framework for accurate interpretation of research findings related to motor performance.

Main Methods:

• Statistical re-examination of the relationships between CE, AE, and VE.
• Analysis under the assumption of a normal distribution for performance data.
• Illustrative examples to demonstrate potential misinterpretations of AE.

Main Results:

• Absolute error (AE) is statistically dependent on and predictable from algebraic error (CE) and within-subject variance (VE).
• The information conveyed by AE is fully contained within CE and VE, depending on their ratio.
• Specific conditions (CE/VE ratio) dictate whether CE, VE, or a combination of both best represents the error information.

Conclusions:

• AE provides no unique information beyond that contained in CE and VE.
• Understanding the interplay between CE and VE is crucial for accurate performance error assessment.
• This research offers a more precise statistical foundation for interpreting motor performance data.