What are the standard ways to fight "observer bias" ?
What are the standard ways to fight "observer bias" ?
Answer: The first thing that we do to fight this is to always use multiple observers and make sure that they agree very strongly about what happened; we require an inter- coder reliability of .90 or better (which is the highest required correlation in psych methods). We also use tricks to limit the options of the observers and prevent overload, such as checklists, time-window sampling, and event-triggered sampling.
Comments
Post a Comment