COMM 88 Chapter Notes - Chapter 7: Field Experiment, Stereotype Threat, Internal Validity
Chapter 7
●Experiments provide ideal model for testing hypotheses and inferring causal relationships
○To determine association
■experiments compare outcomes in two or more groups or conditions
○To establish direction of influence
■The manipulated
IV (the cause) always precedes the measured
DV (the effect)
○To control for variables that might produce a spurious association
■Participants are randomly assigned to conditions
○Everything except the experimental manipulation must stay the same throughout the
entire experiment
●Basic features of experiments
○Hypotheses concern the relationship between two variables
○At least 2 groups or conditions, represented by the categories of the IV
○All groups treated exactly alike
○Participants are randomly assigned to one group or the other
●certain types of empirical evidence are regarded as essential for causal statements
1. Association
a. Evidence that X and Y vary together in a way predicted by hypothesis
2. Direction of influence
a. Evidence that X affected Y rather than Y affected X
3. The elimination of plausible rival explanations
a. Evidence that one or more variables other than X did not cause the observed
change in Y
●Statistical significance - likelihood or probability that an association is due to random process
●Posttest-only control group design - the most basic experimental design in which the dependent
variable is measured after the experimental manipulation
●Pretest-posttest control group design - an experimental design in which the dependent variable is
measured both before and after the experimental manipulation
●Factorial design - an experiment in which two or more variables (factors) are manipulated
○Given that social events are often caused by a number of variables, it makes sense to
study several possible causes or iVs at the same time
○These IVs are referred to as factors when there are two or more
■In this picture 11 means subject was exposed to factor 1 and than factor 1 again,
12 means factor 1 then 2 and so on
Document Summary
Experiments provide ideal model for testing hypotheses and inferring causal relationships. Experiments compare outcomes in two or more groups or conditions. To control for variables that might produce a spurious association. The manipulated iv (the cause) always precedes the measured dv (the effect) Everything except the experimental manipulation must stay the same throughout the entire experiment. Hypotheses concern the relationship between two variables. At least 2 groups or conditions, represented by the categories of the iv. Participants are randomly assigned to one group or the other: association. Statistical significance - likelihood or probability that an association is due to random process. Posttest-only control group design - the most basic experimental design in which the dependent. Pretest-posttest control group design - an experimental design in which the dependent variable is variable is measured after the experimental manipulation measured both before and after the experimental manipulation. Factorial design - an experiment in which two or more variables (factors) are manipulated.