SOCW 671: #6 Research Designs Review for 1 st Quiz
Research Designs Research Designs: Useful in controlling for internal & external threats to validity. Three main types: Experimental Quasi-experimental Non-experimental
Criteria for Inferring Causality The cause precedes the effect in time The two variables must be empirically correlated with one another The observed empirical correlation between cannot be explained away as being due to the influence of some third variable that causes both of them
Internal and External Validity Internal Validity - the confidence we have that the results of a study accurately depict whether one variable is or is not the cause of another variable. External Validity – the extent to which the causal relationship depicted in a study can be generalized beyond the study conditions.
Threats to Internal Validity History Maturation (passage of time) Testing Instrumentation Statistical Regression (not the kind we talked about last week) Selection Biases Experimental Mortality Ambiguity of Direction of Causal Influence Diffusion or Imitation of Treatment
External Validity and Field Experiments Realism – is the experiment realistic Reactivity – research subjects acting as if they are in an experiment (Hawthorn Effect), novelty effect (intervention wears off), or demand characteristics (when subjects pick up cues of how they are suppose to act) Field Experiments – when research is conducted in the field and not the lab, there is less control
Experimental Focus is on generalizability ANOVA and it’s various forms are the most common statistic. Usually use statistics that evaluate group differences Need two conditions: 1) researcher manipulates independent variables 2) random selection and assignment
Experimental Designs (selected) Posttest Only Pretest-Posttest Solomon Four Group
Quasi-experimental (selected) Quasi-experimental: more common at organizational level. Frequently look at relationships between variables. Work with the population and sample provided.
Quasi-experimental Designs Time series (often used in single subject designs) Nonequivalent Control Group Design
Pre-experimental (Non- experimental) Often just a one-shot deal. Mostly use descriptive statistics One Group Pretest-Posttest Design Static Group Comparison