Psychologists choose between two broad types of research design, experimental and non-experimental, and only experiments, which randomly assign participants to groups and manipulate a variable, can establish cause and effect.
Researchers use non-experimental studies when experiments would be impossible, impractical, or unethical, such as studying childhood trauma.
Without random assignment, groups may differ in many ways besides the variable of interest, so researchers cannot tell which factor caused the differences.
Four non-experimental methods are correlational studies, naturalistic observation, case studies, and meta-analyses.
Case studies give rich detail about one case but may not apply to other people, while meta-analyses combine many studies to reveal patterns no single study shows.
Key terms
experimental design
A study in which participants are randomly assigned to groups and the researcher manipulates a variable; it can establish cause and effect.
non-experimental design
A study without random assignment or manipulation of variables; it cannot establish causation.
random assignment
Using chance to place participants into groups, so the groups don't differ in other systematic ways.
correlational study
A study that measures two variables to see whether they are related, without manipulating either one.
correlation
A relationship between two variables (they are co-related); it does not prove that one caused the other.
naturalistic observation
Watching and recording behavior in its natural setting without interacting with the people being observed.
case study
An in-depth investigation of one person, group, or event.
meta-analysis
A study about studies: analyzing the results of many previously published studies instead of collecting new data.
Watch out
A correlation shows two variables are related, not that one caused the other — only an experiment with random assignment can show causation.