The mechanism
Simple reaction tasks require one known response, while choice tasks require mapping one of several stimuli to one of several responses. More alternatives and greater uncertainty usually lengthen the decision stage. Weak or similar stimuli slow evidence collection, and incompatible key layouts slow response selection. Motor execution remains part of the recorded total even when the scientific question concerns perception or decision.
What to measure
Prefer the median of correct trials to a single best time, and report accuracy, slow-lapse frequency, and within-session variability. Separate simple and choice tasks, preserve the number of alternatives, and log the input device. For symbol–digit tasks, count correct substitutions within the time limit and also retain errors, omissions, and learning across repeated blocks. Across sessions, model the early practice gain separately from later change and compare equivalent difficulty levels.
Useful practice conditions
- Calibrate the display and input path, and avoid comparing raw milliseconds across devices without a latency check.
- Use enough trials to stabilize the median while keeping blocks short enough to limit fatigue.
- Compare users mainly with their own baseline under the same task, device, and accuracy requirement.
Limits of the conclusion
A faster score can reflect familiarity, a changed response strategy, hardware, or motor speed as well as faster task processing. It is not a stand-alone measure of general ability and should not be used as a diagnosis. Claims are strongest when they refer to the exact task, preserve accuracy, and are replicated across more than one session.