Sampling & Validity Flashcards
Random sample
Every person in the population has an equal chance of getting chosen, thus creating generalized results of the intended population.
- In psychology, predictions are usually made from a general level to a specific level, therefore random samples are rarely used.
Non-random sample
Individuals from a highly specialized subpopulation.
Less external validity, since findings might only apply to the subpopulation and cannot necessarily be generalized to the intended population.
Simple random sampling
Randomly selecting a certain number of participants from the population. This method does not guarantee representatives.
Stratified sampling
Divide the population into segments and choose and get a random sample of equal size from each segment. This method might lead to over/under-representation.
Proportionate sampling
The proportion of people in a population is reflected in the sample; e.g. 45% Caucasian, 37% African-American, 18% Hispanic.
Systematic sampling
Sampling every kth element after a random start; e.g. every 5th student on the university enrollment list.
Cluster Sampling
Identifying naturally occurring groups of participants and selecting a specific cluster, e.g. classroom.
Multistage sampling
Related to cluster sampling, but choosing individuals within the clusters rather than entire clusters.
Volunteer bias
Individuals who chose to participate in a study differ from those who didn’t. A sample made up entirely of volunteers is biased which affects external validity.
Internal validity
The ability of research design to adequately test the hypothesis it was designed to test.
External validity
Results from the study can be extended (generalized) beyond the research setting and sample in which they were obtained. However, that might not always be the aim of psychology.
Threats to internal validity
- History: events may occur between observations.
- Maturation: performance change due to age/fatigue.
- Testing: pre-test sensitizes participants to study aim.
- Instrumentation: changes in observer/instrument.
- Statistical regression: outliers move closer to the mean.
- Biased selection of subjects.
- Experimental mortality: differential loss of participants.
Threats to external validity
- Reactive testing: pre-test affects participants’ response.
- Participation bias and the independent variable.
- Reactive effects of experimental arrangements.
- Multiple treatment inferences: several parts of study.
Lab setting
+ Control of important variables.
+ Approach is more ethically correct.
- Specialized > generalized
- The real-world situation might not be authentic.
- Realism; participants’ behavior has consequences.
Internal validity > external validity
Field setting
+ Participants natural environment; authentic results.
+ Results can easily be generalized.
- Limited control over confounding variables.
Internal validity < external validity
Factors to consider when acquiring human participants for research.
- Research setting: lab/field research.
- Special needs: screening for specific characteristics.
- Ethics: APA guidelines + federal and state laws.
