Three Kinds of Methods1. Observational2. Archival3. SurveysHowever, the most important distinction is that between correlational research and experimental research.Observational ResearchInvolves looking at a phenomenon in a somewhat systematic wayUsually includes taking notes and interviewing peopleCan be used to get ideas/generate hypotheses about social behavior (e.g., “Dissonance Theory”)Not able to test ideas (no careful measurement of variables)Archival ResearchInvolves using past records to look for relationships between variablesNewspapers, police reports, hospital records, and so onCan be used to test theories about social behavior(e.g., “Culture of Honor” – Cohen et al., 1996)Example: Police and FBI records were used to test a theory that personal reputation was more important in the southern United States than in the northern United StatesHomicides over insults are more common in the South than the NorthWay to remember: If the data already exists and the public is able to get access to it, then it is probably archival (e.g. freely available public records)Correlational ResearchCorrelational Research: Research that examines the (linear) relationship between variables without assigning participants to different situations or conditions; pretty much studies the relationship between two thingsCorrelational value r can only be between -1 and 1.The farther r is from 0, the stronger the relation between the variablesIn other words the absolute value of the correlation determines the strengthA correlation of 0 indicates no relation.r=undefined when the line is a straight horizontal lineThe sign (either positive or negative) indicates the type/direction of the relationPositive Correlation: the two variables move in the same directions (when one goes up the other goes up; when one goes down, the other goes down)Negative Correlation: the two variables move in opposite directions (when one goes up the other goes down; when one goes down the other goes up)Graphs: read them from left to rightpositive correlations go upnegative correlations go downCorrelation sizeSmall(+) .10 to .29(-) -.29 to -.10Medium(+) .30 to .49(-) -.49 to -.30Large(+) .50 to 1.00(-) -1.00 to -.50Properties of correlational researchExamines the strength of the relationship, if any, between variablesResearcher measures the variables, but does NOT manipulate or control them – measure them as they naturally occurAllows prediction of one variable, if you know the value of the other variableCausation definedCause and effect have to be relatedCause has to precede the effectDon’t know often—usually two variables are simply measuredCan make assumptions (e.g. gender and aggression)Other explanations of the cause-effect relationship have to be eliminatedDifficult to accomplish. Requires the measuring and controlling of all other variablesLimits of CorrelationsDifficult/Impossible to make inferences about causalityCorrelation does usually imply causation!Reverse causality problem: Does Y cause X?Third Variable problem: When some unmeasured third variable is causing one or both of the measured variablesSelf-selection: when the participant selects his or her level on each variable, bringing with this value unknown other properties that make causal interpretation difficult.is one big reason why you cannot rule out other explanations and say it is definite X that causes YThe people who do the behavior you are interested in, may also being doing other behaviors that go along with it. Those other behaviors could be responsible for the result.Example: people on diets may also exercise, go to doctors more often, have better standards of livingSolving the limitations of correlational researchCorrelational research has two limitations1. Temporal PrecedenceNeed to know that X came before YTherefore, researchers need to manipulate a variable to establish temporal precedence—have people receive a one level of a variable and have other people receive another level2. Self SelectionNeed to know that the only difference between the groups is a hypothesized causeRandomly assign people to the manipulation (e.g. flip a coin)Experimental ResearchExperimental research: (Think of it like a Medicine Drug Trial)Research that involves assigning participants to different situations or conditions.Participants should be randomly assigned to different conditionsConditions are controlled or manipulated by the researcherBehaviors are systematically measuredComparisons of how different manipulations affect behavior allow researchers to determine causal influences of behaviorExperimental MethodsIndependent variable: the variable that is manipulated by the researcherThe independent variable is hypothesized to cause changes in the dependent variablesDependent variable: The variable that is measuredOften a change in behavior, feelings, or evaluationControl condition: A condition identical to the experimental condition but absent from the independent variableUsed to compare the effectiveness of the variablesRandom assignment: Random assignment to conditions ensures that individual differences are evenly distributed across conditionsCan infer that differences between experimental and control groups are due to the experimental manipulation and not to differences between the types of people that were in each conditionCan solve self-selection problemEvaluating Measurements/ Scales/ TestsScales/Psychological Measurements tend to evaluated along two criteria:1. Reliability: How consistently a test will measure the variable of interestIf you took the same test twice, would it give you the same score?Aka consistency2. Validity (Measurement Validity): The degree that a test accurately measures the variable of intereste.g. are IQ tests true measures of what people think of as intelligence?Aka accuracyMethods of Social Psych 08/30/2012Three Kinds of Methods1. Observational2. Archival3. Surveys-However, the most important distinction is that between correlational research and experimental research.Observational Research• Involves looking at a phenomenon in a somewhat systematic way • Usually includes taking notes and interviewing people• Can be used to get ideas/generate hypotheses about social behavior (e.g., “Dissonance Theory”)• Not able to test ideas (no careful measurement of variables)Archival Research-Involves using past records to look for relationships between variablesoNewspapers, police reports, hospital records, and