PSYCO212 Lecture Notes - Lecture 33: Construct Validity, Internal Validity, Experiment

• School: University of Alberta
• Department: Psychology
• Course: PSYCO212
• Professor: Poliana Barbosa

Document Summary

Quasi-experiments enable researchers to take advantage of real-world opportunities to study interesting phenomena and events. These designs can enhance external validity; the likelihood that the patterns observed will generalize to other settings and individuals. Many interests of the researchers would be unethical to perform in a true experiment, and therefore quasi-experiments are used. Often, quasi-experiments show exceed the construct validity for the independent variable. We also need to ask how well the dependent variables are measured. Statistical validity: we would ask how large the group differences were (effect size) and whether the results were statistically significant. When a quasi-experiment uses a between-groups design, the groups in the study can look similar to those in correlational studies. Quasi-experiments and correlational designs also have similar internal validity concerns. Whereas correlational researchers primarily measure variables in a sample and analyze their relationships, quasi-experimental researchers more actively select groups for an independent variable so they can achieve greater degree of internal validity.