Logarithmic Versus Linear Change in Step Size When Using an Adaptive Threshold-Seeking Procedure in a Frequency...

Area of Science:

• Auditory perception research
• Psychoacoustics
• Signal detection theory

Background:

• Adaptive threshold-seeking procedures often employ different step-size rules (logarithmic, linear) without clear justification.
• Hypothesis: Linear rules may improve accuracy for poorer performers due to fixed step sizes, unlike variable logarithmic steps.

Purpose of the Study:

• To investigate the impact of logarithmic versus linear step-size rules on frequency discrimination (FD) thresholds.
• To compare the efficacy of these rules across different age groups and experimental designs.

Main Methods:

• Three experiments were conducted with human participants (young adults and children).
• Methods included 3-interval 2-alternative forced choice (3I2AFC) and 2-interval 2AFC tasks.
• Monte Carlo computer simulations were used to validate empirical findings.

Main Results:

• No significant differences in FD thresholds were observed between logarithmic and linear rules across all conditions.
• Computer simulations corroborated empirical data, showing comparable thresholds for both rules in most simulations.
• Linear rules showed potential for higher accuracy but also produced more outliers, particularly with decreased participant attention.

Conclusions:

• The choice of step-size rule (logarithmic vs. linear) has minimal impact on frequency discrimination thresholds.
• Outlier reduction may be achieved by monitoring and ensuring participant attention at the start of threshold estimation runs.