Signal-detection analyses of conditional discrimination and delayed matching-to-sample performance

Area of Science:

• Behavioral science
• Cognitive psychology
• Animal behavior

Background:

• Quantitative analyses in conditional discrimination and delayed matching-to-sample procedures face challenges when subjects exhibit no errors.
• This lack of errors complicates the analysis of stimulus control and reinforcer control.

Purpose of the Study:

• To examine two prevalent methods used to address data analysis when no errors occur in behavioral tasks.
• To evaluate the impact of these methods on the accuracy of quantitative analyses.

Main Methods:

• Monte Carlo simulations were employed to model performance in these tasks.
• The simulations assessed the outcomes of two distinct data analysis techniques applied to error-free performance.

Main Results:

• Both examined methods introduced systematic deviations into the analytical results.
• The use of these methods carries a significant risk of generating misleading conclusions.

Conclusions:

• Standard methods for analyzing error-free data in conditional discrimination and delayed matching-to-sample tasks can distort findings.
• Researchers must exercise caution to avoid erroneous conclusions regarding behavioral models, signal detection, and animal short-term memory.