Averaging learning curves across and within participants

Area of Science:

• Cognitive psychology
• Learning sciences
• Data analysis

Background:

• Averaged learning curves are commonly used to summarize group learning trends.
• Concerns exist that averaging may distort the representation of individual learning trajectories.
• Standard averaging methods can obscure important variations in learning.

Purpose of the Study:

• To investigate the impact of different averaging techniques on learning curve analysis.
• To determine if averaged learning curves accurately reflect individual learning patterns.
• To identify averaging methods that preserve functional form and reduce noise.

Main Methods:

• Analysis of practice curve data across various experimental paradigms.
• Comparison of arithmetic and geometric averaging over participants.
• Evaluation of block averaging and other smoothing techniques on individual curves.
• Assessment of the distortion of power and exponential function fits.

Main Results:

• Arithmetic averaging of learning curves distorts individual learning, particularly for power and exponential functions.
• Geometric averaging also fails to consistently avoid distortion.
• Block averaging and similar smoothing methods minimize distortion of functional form.
• These methods retain the noise-reduction benefits of averaging.

Conclusions:

• Averaging learning curves requires careful consideration of the method used.
• Block averaging is a preferred method for analyzing learning curves to avoid distortion.
• Understanding averaging effects is crucial for accurate interpretation of learning data.