Analyzing coefficients of psychophysical power functions

Area of Science:

• Mathematical modeling
• Psychophysics
• Statistical analysis

Background:

• Power functions are used to model relationships between physical stimuli and responses.
• Coefficients of these functions, including intercepts and exponents, can be influenced by measurement units.

Purpose of the Study:

• To analyze the mathematical properties of power function coefficients.
• To investigate how the choice of measurement units for physical stimuli affects correlations between intercepts and exponents.
• To examine the impact of stimulus geometric mean on statistical test power.

Main Methods:

• Analysis of mathematical properties of power function coefficients.
• Correlation analysis between intercepts (logarithm of coefficients) and exponents.
• Investigation of the geometric mean of stimulus measures.
• Evaluation of the power of statistical tests for differences between mean intercepts.

Main Results:

• The correlation between intercepts and exponents is dependent on the unit of measurement for physical stimuli.
• When stimulus geometric mean deviates from one, the absolute correlation size increases.
• Positive correlation occurs when geometric mean is less than one; negative when greater than one.
• Statistical test power is reduced as the geometric mean deviates from one.

Conclusions:

• The choice of measurement units significantly influences the interpretation of power function relationships.
• Researchers must consider the geometric mean of stimuli when analyzing power function data and conducting statistical tests.
• Deviations from a geometric mean of one can attenuate statistical power and alter observed correlations.