On the Utility of Integrated Speed-Accuracy Measures when Speed-Accuracy Trade-off is Present

Area of Science:

• Cognitive psychology
• Psychometric methods

Background:

• Integrated measures of speed and accuracy are used to simplify data analysis.
• Claims exist that some integrated measures are insensitive to speed-accuracy trade-off (SAT).

Purpose of the Study:

• To systematically examine the sensitivity of established integrated measures to SAT.
• To evaluate the effectiveness of different experimental designs in mitigating SAT confounds.

Main Methods:

• Four simulation studies were conducted using five established integrated measures.
• Designs included repeated measures crossing an experimental factor with SAT settings.
• SAT was modeled using drift-diffusion and two-phase models, with balanced trade-offs in one study.

Main Results:

• Only balanced trade-offs showed two measures insensitive to SAT.
• In other contexts, all measures were sensitive to SAT.
• Mixed conditions designs ensured all integrated measures detected the experimental variable's effect.

Conclusions:

• Most integrated measures are sensitive to SAT.
• A mixed conditions repeated measures design effectively neutralizes SAT effects.
• This design allows for reliable detection of experimental variables despite SAT.