Ellen Katherine Rothschild
Chapter 2: The Methods of Social Psychology
Why do social psychologists do research (and why should you want to read about it?)
Hindsight bias: people’s tendency to be overconfident about whether they could have predicted a given
• Social psychological research teaches people how to interpret and predict the outcomes of various social
experiences and helps them understand their own behavior and that of others.
How do Social Psychologists test ideas?
Hypothesis: a prediction about what will happen under particular circumstances
• Often test broader theories about behavior
o Theory: a body of related propositions intended to describe some aspect of the world.
Observational research: research can be merely a matter of looking at a phenomenon in some reasonably
systematic way with a view to understanding what is going on and coming up with hypothesis about why things
are happening as they are.
• Participant observation: involves observing some phenomenon at close range
• Often observe in a semi formal way, taking notes and interviewing participants
Archival research: researchers look at evidence found in archives of various kinds – record books, police
reports, sports statistics, newspaper articles, and databases containing anthropological descriptions of people in
Surveys: simply asking people questions (interviews or written questionnaires)
• It is important that the people in the survey be a random sample of the population as a whole: give
everyone in the population an equal chance of being selected.
• Convenience sample: not random and may be biased – may include too many of some kind of person
and too few of another. (taken by convenience i.e. they are easily accessible)
• The number of people you need in a sample is independent of population size.
Correlational research: research that does not involve random assignment to different situations, or conditions,
and that psychologist conduct just to see whether is a relationship between the variables.
• Experimental research: in social psychology, research that randomly assigns people to different
conditions, or situations, and that enables researchers to make strong inferences about how these
different conditions affect people’s behavior.
Correlation does not equal causation!
• Reverse causation: when 1 variable is assumed to cause variable 2, yet the opposite direction may be
• Third variable: when variable 1 does not cause variable 2 and variable 2 does not cause variable 1, but
rather some other variable exerts a causal influence on both.
• Self selection: a problem that arises when the participant, rather than the investigator, selects his or her
level on each variable, bringing with this value unknown properties that make casual interpretation of a
o Correlational research cannot prove causation because of self selection
• Can only look at the degree of relationship between two or more variables. (range of 0 to 1) • Correlational studies can be very helpful in alerting investigators to various possibilities for valid causal
hypotheses about the nature of the world.
• Sometimes the best option when an experimental study would be difficult or unethical
• If correlational studies cannot prove that a causal relationship exists, clever analysis of correlational
data can be quite persuasive abut the meaning of a relationship
Longitudinal study: a study conducted over a long period of time with the same population, which is
periodically assessed regarding a particular behavior.
The best way to be sure about causality is to conduct an experiment.
• Independent variable: in experimental research, the variable that is manipulated; it is hypothesized to be
the cause of a particular outcome.
• Dependent variable: in experimental research, the variable that is measures (as opposed to manipulated);
it is hypothesized to be affected by manipulation of the independent variable.
• The great power of experiments comes from their ability to expose participants to different levels of the
independent variable by random assignment: assigning participants in experimental research to different
groups randomly, such that they are as likely to be assigned to one condition as another.
o Rules out the possibility of self-selection biases in sam