RESEARCH DESIGN : 1
Kinds of support for making CAUSAL interpretations of observed relationships quality of theory research design used measurement procedures
RESEARCH DESIGN the strategy used to collect the data needed to test our hypotheses some kinds of research designs (or data collection strategies) are more causally convincing than others
CONDITIONS that must be met to establish claim that one variable is the cause of another must demonstrate a statistical relationship between the two presumed cause must occur BEFORE the presumed effect researcher must rule out alternative causes (RIVAL HYPOTHESES)
INTERNAL VALIDITY The ability of a research design to help us to rule out rival hypotheses
Threats to Internal Validity NOT X > Y BUT Y > X 1. REVERSE CAUSATION
TIME THREATS : Changes in the outcome variable measured over time within subjects due to factors other than the independent variable. history maturation testing reactivity instrumentation
GROUP THREATS: Differences in outcomes due to differences between groups other than “treatment” differences nonequivalence mortality regression (toward the mean)
Designs that are vulnerable to regression toward the mean threat
TRUE EXPERIMENT: to achieve internal validity SELECT SUBJECTS from some population--goal: representativeness ASSIGN SUBJECTS to groups or treatment conditions. How??? Why??? PRETEST---optional---Why not??? TREATMENT POST-TEST
In Dooley’s diagram notation X X R R In the absence of a pre-test
The key to enhancing internal validity via an experimental research design is CONTROL control over the independent variable which is a treatment introduced and manipulated control over group membership. Subjects are assigned to groups. control over the setting. Other circumstances are held constant
Remaining sources of contamination??? experimenter expectancy
Examples of attempts to apply true experiment research design to social phenomena??? Housing and Employment Audit Studies