Uncontrolled environments
Participants may take tests while tired, distracted or using different devices. These differences can affect accuracy and response time.
Practice effects
People may improve because they become familiar with task types. A tracking system should account for this instead of treating every improvement as cognitive change.
Calibration takes time
Percentiles and comparison groups require a sufficiently large dataset. Early versions should avoid presenting scores with false precision.
Not clinical diagnosis
Online cognitive tests should not claim to diagnose dementia, Alzheimer’s disease, ADHD, depression or any medical condition unless they are clinically validated for that use.