Statistical significance testing with mahalanobis distance for thresholds estimated from constant stimuli method

Area of Science:

• Psychology
• Statistics
• Psychophysics

Background:

• Traditional statistical tests like t-tests and ANOVA are unsuitable for analyzing perceptual thresholds derived from psychometric functions.
• Existing methods cannot derive necessary statistics for significance testing of individual observer thresholds estimated via the constant stimuli method.

Purpose of the Study:

• To introduce a novel statistical testing method for assessing the significance of threshold differences estimated using the constant stimuli method.
• To extend the application of this new method to analyze main effects and interactions in multifactor experiments.

Main Methods:

• The proposed method leverages the asymptotic normality of maximum likelihood estimators.
• It utilizes the properties of multivariate normal distributions to analyze threshold differences.
• The method is applicable to thresholds estimated from constant stimuli, adjustment, and staircase methods.

Main Results:

• The new method successfully assesses the statistical significance of differences between thresholds.
• It demonstrates capability in evaluating main effects and interactions in multifactor experiments.
• Simulation data confirmed the method's assumptions, power, and type I error rates.

Conclusions:

• The developed statistical method provides a valid approach for analyzing perceptual thresholds from the constant stimuli method.
• This technique offers a powerful alternative to conventional statistical analyses in psychophysics.
• The method's versatility extends to various threshold estimation techniques and experimental designs.