External validity concerns how far the results of a study can be generalized.
For example suppose the study at a "fine dining" restaurant (like Ruth's Chris Steak House in Louisville, KY) found that servers who introduced themselves by name got higher tips (23%) than those who did not (15%). The E concluded, " This shows personalizing service leads to higher customer satisfaction."
Now consider each of the external validity questions.
(a) External validity would exist for settings if the same result occurs in other places where tips are given. For example, would the same results occur in fine dining restaurants besides Ruth's; in other restaurant types like casual dining; country clubs; or even other settings where tips are given like to hotel maids, taxi drivers, baggage handlers and so on. The more of these different settings the same results are found in the stronger the evidence for the external validity of the finding across settings.
b) Populations means the type of people who participated in the study, the restaurant patrons. If people from different regions of the country; of different socioeconomic background; educational level; age; and so on show the same higher tipping pattern for servers who used name introductions the stronger the evidence for external validity for the population variable.
c) Treatment variables concerns what other types of personalization besides name introduction would work. For example; if the server asked the customer's name; the service worker commented on some personal item of the customer (e.g. "That's a neat watch."); or some aspect of the customer's preference (e.g. "If you like the pumpkin soup you would probably like the squash side dish.") and so on and those different types of personalization comments led to higher tips the finding would demonstrate external validity for treatment variables.
d) Measurement variables concern customers showing other signs of liking the personalized service besides giving higher tips. For example; if they ordered more items; visited the business more frequently; recommended it to their friends; and so on the findings would show external validity for measurement variables..
Remember it is first essential that the study have internal validity. If it does, then you can ask these external validity questions (which may lead you to look up other similar studies) to see how far the study results can be generalized. Research programs in psychology (e.g. Do 12 step programs like Alcoholics Anonymous work for other addiction problems?) are often based on doing more studies to test external validity.
In summary, external validity is for what other (a) settings, (b) populations, (c) treatment variables and (d) measurement variables can the same results be found? Practically, external validity is how and where else can the study be applied. Any one study is limited in the range of these possible variables it includes. Operational definition is the term for how a general concept is represented in any particular study. For example, in Osborn (1996) the setting was "small groups in a college classroom"; the population, the participants, were "male and female college students 18-21 years old enrolled in an introductory psychology class"; the treatment (makeup versus no-makeup independent variable) was represented by "10 color photos taken from a cosmetics ad of models either wearing makeup or not." The measurement variable, the dependent variable, physical attractiveness, was defined as a "rating from 1 (extremely attractive) through 8 (extremely unattractive)." Obviously all these factors could have been defined somewhat differently. To the extent other studies are carried out, varying these factors and coming up with the same results, we have confidence in the strength of the finding.
Return to Home