Combining speed and accuracy to control for speed-accuracy trade-offs(?)

Area of Science:

• Cognitive Psychology
• Psychometrics

Background:

• Psychological experiments often require participants to balance speed and accuracy.
• Variations in interpreting these instructions can lead to unpredictable effects in reaction times (RTs) and error rates.
• Differential speed-accuracy trade-offs can introduce spurious experimental effects.

Purpose of the Study:

• To evaluate the performance of different scores used to combine reaction time and accuracy.
• To determine which score is most robust against confounding speed-accuracy trade-offs.

Main Methods:

• Simulated data from the standard diffusion model were analyzed.
• The inverse efficiency score (IES), rate-correct score (RCS), and linear-integrated speed-accuracy score (LISAS) were compared.
• The balanced integration score (BIS) was evaluated for its sensitivity to speed-accuracy trade-offs.

Main Results:

• IES, RCS, and LISAS demonstrated unequal weighting of speed and accuracy, varying with accuracy levels.
• These scores were found to be highly sensitive to speed-accuracy trade-offs.
• BIS showed relative insensitivity to speed-accuracy trade-offs while preserving genuine experimental effects.

Conclusions:

• Existing scores like IES, RCS, and LISAS are susceptible to confounding by speed-accuracy trade-offs.
• The balanced integration score (BIS) offers a more reliable measure by integrating speed and accuracy with equal weights.
• BIS is recommended for psychological research to mitigate spurious findings due to varying speed-accuracy strategies.