6.5 External Validity

A study’s external validity is related to its reproducibility.  A study that lacks external validity cannot be applied to other people, and other contexts, beyond the scope of the original experiment. 

All of the aforementioned internal validity threats can influence external validity as well.

One common threat to external validity is the use of a non-representative sample.  Many experiments are performed on college campuses, because researchers have a readily-available pool of subjects to draw from when working with students.  While there is nothing automatically wrong with this, they should always keep in mind that a sample of college students is not representative of the overall population, in terms of age, health, priorities, income, wealth, etc.  Other threats to external validity include the behavior of researchers (is the questioner showing tacit approval or disapproval of particular answers through his body language?), the setting/environment (does the blue paint on the wall in a room tend to make people feel calmer?), and the use of pre-test questionnaires (do questions about time spent with family make people more likely to support work-from-home arrangements for employees?)