External Validity and Generalizability
Claims and Validity
Claims from research are invalid if the research itself is invalid.
Validities assess how well research is conducted.
External Validity
External validity assesses how well research results apply outside the study's specific context (participants, settings, situations).
The goal of most research is for the results to apply to the real world, not just the lab.
Poor external validity means the research doesn't generalize to real-world behavior.
Terms synonymous with external validity:
Generalizability: How general are the results?
Ecological validity: Would the same results occur in a real-world setting?
Aspects of External Validity
Sample and Population:
The tested group should represent the larger group you want to learn about.
A sample is a subset of the population.
Sample Size:
Need to test enough subjects to be representative.
Statistical power analysis can determine the necessary number of participants.
Sample Composition:
The sample should include a variety of people representative of the full population.
Consider factors like region, race, and gender.
College Sophomore Problem:
Much psychological research relies on college freshmen/sophomores.
This is mainly due to their easy accessibility at universities.
College students may not represent the broader population due to age, education level, etc.
Harlow's Monkey Experiments and External Validity
Do the behaviors of Harlow's monkeys represent the behavior of all monkeys or monkeys in the wild?
Does the behavior of Harlow's monkeys represent the behavior of humans?
It's assumed that monkeys are close enough to humans to generalize, but this may not be the case.
Real World Example: Colombian Peace Deal Vote
A peace deal was put to a national vote but was narrowly rejected.
Low voter turnout and a hurricane potentially skewed the sample, making it less representative.
This highlights how important a representative sample is for accurate results.
In the context of elections, the 'population' refers to the entire group of eligible voters, while a 'sample' is the group of people who actually vote (