Treatment Drift occurs when there is a change in the way the independent variable (treatment or intervention) is applied during later phases of an experiment compared to how it was implemented at the beginning. This can affect the validity of the study by introducing inconsistency in the intervention process.
Example
In a study where teachers use a token economy to reinforce student behavior, treatment drift might occur if, over time, the teachers start giving tokens for different behaviors than they initially reinforced, or they apply the system less consistently. This variation in how the independent variable is delivered can lead to inaccurate conclusions about the effectiveness of the treatment.